xxq250
/
homework-jianmu
1
0
Fork 0
Code Issues Pull Requests 1 Actions Packages Projects Releases Wiki Activity

Merge branch '3.0' into cpwu/3.0

This commit is contained in:
cpwu 2022-07-22 14:36:00 +08:00
parent cd0e66a694 0c33cdeefb
commit d55d195e8a
205 changed files with 7036 additions and 5811 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
cmake/cmake.options
View File

@ -51,6 +51,13 @@ IF(${TD_WINDOWS})
"If build unit tests using googletest"
ON
)
option(
TDENGINE_3
"TDengine 3.x"
ON
)
ELSEIF (TD_DARWIN_64)
add_definitions(-DCOMPILER_SUPPORTS_CXX13)
option(

2
docs/zh/02-intro.md
View File

@ -52,7 +52,7 @@ TDengine的主要功能如下：
采用 TDengine，可将典型的物联网、车联网、工业互联网大数据平台的总拥有成本大幅降低。表现在几个方面：
1. 由于其超强性能，它能将系统需要的计算资源和存储资源大幅降低
2. 因为采用 SQL 接口，能与众多第三放软件无缝集成，学习迁移成本大幅下降
2. 因为采用 SQL 接口，能与众多第三方软件无缝集成，学习迁移成本大幅下降
3. 因为其 All In One 的特性，系统复杂度降低，能降研发成本
4. 因为运维维护简单，运营维护成本能大幅降低

1
docs/zh/27-train-faq/03-docker.md → docs/zh/05-get-started/01-docker.md
View File

@ -1,4 +1,5 @@
---
sidebar_label: Docker
title: 通过 Docker 快速体验 TDengine
---

296
docs/zh/05-get-started/03-package.md Normal file
View File

@ -0,0 +1,296 @@
---
sidebar_label: 安装包
title: 使用安装包立即开始
---
import Tabs from "@theme/Tabs";
import TabItem from "@theme/TabItem";
:::info
如果您希望对 TDengine 贡献代码或对内部实现感兴趣，请参考我们的 [TDengine GitHub 主页](https://github.com/taosdata/TDengine) 下载源码构建和安装.
:::
TDengine 开源版本提供 deb 和 rpm 格式安装包，用户可以根据自己的运行环境选择合适的安装包。其中 deb 支持 Debian/Ubuntu 及衍生系统，rpm 支持 CentOS/RHEL/SUSE 及衍生系统。同时我们也为企业用户提供 tar.gz 格式安装包。
## 安装
<Tabs>
<TabItem value="apt-get" label="apt-get">
可以使用 apt-get 工具从官方仓库安装。
**安装包仓库**
```
wget -qO - http://repos.taosdata.com/tdengine.key | sudo apt-key add -
echo "deb [arch=amd64] http://repos.taosdata.com/tdengine-stable stable main" | sudo tee /etc/apt/sources.list.d/tdengine-stable.list
```
如果安装 Beta 版需要安装包仓库
```
echo "deb [arch=amd64] http://repos.taosdata.com/tdengine-beta beta main" | sudo tee /etc/apt/sources.list.d/tdengine-beta.list
```
**使用 apt-get 命令安装**
```
sudo apt-get update
apt-cache policy tdengine
sudo apt-get install tdengine
```
:::tip
apt-get 方式只适用于 Debian 或 Ubuntu 系统
::::
</TabItem>
<TabItem label="Deb 安装" value="debinst">
1、从官网下载获得 deb 安装包，例如 TDengine-server-2.4.0.7-Linux-x64.deb；
2、进入到 TDengine-server-2.4.0.7-Linux-x64.deb 安装包所在目录，执行如下的安装命令：
```
$ sudo dpkg -i TDengine-server-2.4.0.7-Linux-x64.deb
(Reading database ... 137504 files and directories currently installed.)
Preparing to unpack TDengine-server-2.4.0.7-Linux-x64.deb ...
TDengine is removed successfully!
Unpacking tdengine (2.4.0.7) over (2.4.0.7) ...
Setting up tdengine (2.4.0.7) ...
Start to install TDengine...
System hostname is: ubuntu-1804
Enter FQDN:port (like h1.taosdata.com:6030) of an existing TDengine cluster node to join
OR leave it blank to build one:
Enter your email address for priority support or enter empty to skip:
Created symlink /etc/systemd/system/multi-user.target.wants/taosd.service → /etc/systemd/system/taosd.service.
To configure TDengine : edit /etc/taos/taos.cfg
To start TDengine : sudo systemctl start taosd
To access TDengine : taos -h ubuntu-1804 to login into TDengine server
TDengine is installed successfully!
```
</TabItem>
<TabItem label="RPM 安装" value="rpminst">
1、从官网下载获得 rpm 安装包，例如 TDengine-server-2.4.0.7-Linux-x64.rpm；
2、进入到 TDengine-server-2.4.0.7-Linux-x64.rpm 安装包所在目录，执行如下的安装命令：
```
$ sudo rpm -ivh TDengine-server-2.4.0.7-Linux-x64.rpm
Preparing... ################################# [100%]
Updating / installing...
1:tdengine-2.4.0.7-3 ################################# [100%]
Start to install TDengine...
System hostname is: centos7
Enter FQDN:port (like h1.taosdata.com:6030) of an existing TDengine cluster node to join
OR leave it blank to build one:
Enter your email address for priority support or enter empty to skip:
Created symlink from /etc/systemd/system/multi-user.target.wants/taosd.service to /etc/systemd/system/taosd.service.
To configure TDengine : edit /etc/taos/taos.cfg
To start TDengine : sudo systemctl start taosd
To access TDengine : taos -h centos7 to login into TDengine server
TDengine is installed successfully!
```
</TabItem>
<TabItem label="tar.gz 安装" value="tarinst">
1、从官网下载获得 tar.gz 安装包，例如 TDengine-server-2.4.0.7-Linux-x64.tar.gz；
2、进入到 TDengine-server-2.4.0.7-Linux-x64.tar.gz 安装包所在目录，先解压文件后，进入子目录，执行其中的 install.sh 安装脚本：
```
$ tar xvzf TDengine-enterprise-server-2.4.0.7-Linux-x64.tar.gz
TDengine-enterprise-server-2.4.0.7/
TDengine-enterprise-server-2.4.0.7/driver/
TDengine-enterprise-server-2.4.0.7/driver/vercomp.txt
TDengine-enterprise-server-2.4.0.7/driver/libtaos.so.2.4.0.7
TDengine-enterprise-server-2.4.0.7/install.sh
TDengine-enterprise-server-2.4.0.7/examples/
...
$ ll
total 43816
drwxrwxr-x 3 ubuntu ubuntu 4096 Feb 22 09:31 ./
drwxr-xr-x 20 ubuntu ubuntu 4096 Feb 22 09:30 ../
drwxrwxr-x 4 ubuntu ubuntu 4096 Feb 22 09:30 TDengine-enterprise-server-2.4.0.7/
-rw-rw-r-- 1 ubuntu ubuntu 44852544 Feb 22 09:31 TDengine-enterprise-server-2.4.0.7-Linux-x64.tar.gz
$ cd TDengine-enterprise-server-2.4.0.7/
$ ll
total 40784
drwxrwxr-x 4 ubuntu ubuntu 4096 Feb 22 09:30 ./
drwxrwxr-x 3 ubuntu ubuntu 4096 Feb 22 09:31 ../
drwxrwxr-x 2 ubuntu ubuntu 4096 Feb 22 09:30 driver/
drwxrwxr-x 10 ubuntu ubuntu 4096 Feb 22 09:30 examples/
-rwxrwxr-x 1 ubuntu ubuntu 33294 Feb 22 09:30 install.sh*
-rw-rw-r-- 1 ubuntu ubuntu 41704288 Feb 22 09:30 taos.tar.gz
$ sudo ./install.sh
Start to update TDengine...
Created symlink /etc/systemd/system/multi-user.target.wants/taosd.service → /etc/systemd/system/taosd.service.
Nginx for TDengine is updated successfully!
To configure TDengine : edit /etc/taos/taos.cfg
To configure Taos Adapter (if has) : edit /etc/taos/taosadapter.toml
To start TDengine : sudo systemctl start taosd
To access TDengine : use taos -h ubuntu-1804 in shell OR from http://127.0.0.1:6060
TDengine is updated successfully!
Install taoskeeper as a standalone service
taoskeeper is installed, enable it by `systemctl enable taoskeeper`
```
:::info
install.sh 安装脚本在执行过程中，会通过命令行交互界面询问一些配置信息。如果希望采取无交互安装方式，那么可以用 -e no 参数来执行 install.sh 脚本。运行 `./install.sh -h` 指令可以查看所有参数的详细说明信息。
:::
</TabItem>
</Tabs>
:::note
当安装第一个节点时，出现 Enter FQDN：提示的时候，不需要输入任何内容。只有当安装第二个或以后更多的节点时，才需要输入已有集群中任何一个可用节点的 FQDN，支持该新节点加入集群。当然也可以不输入，而是在新节点启动前，配置到新节点的配置文件中。
:::
## 启动
安装后，请使用 `systemctl` 命令来启动 TDengine 的服务进程。
```bash
systemctl start taosd
```
检查服务是否正常工作：
```bash
systemctl status taosd
```
如果服务进程处于活动状态，则 status 指令会显示如下的相关信息：
```
Active: active (running)
```
如果后台服务进程处于停止状态，则 status 指令会显示如下的相关信息：
```
Active: inactive (dead)
```
如果 TDengine 服务正常工作，那么您可以通过 TDengine 的命令行程序 `taos` 来访问并体验 TDengine。
systemctl 命令汇总：
- 启动服务进程：`systemctl start taosd`
- 停止服务进程：`systemctl stop taosd`
- 重启服务进程：`systemctl restart taosd`
- 查看服务状态：`systemctl status taosd`
:::info
- systemctl 命令需要 _root_ 权限来运行，如果您非 _root_ 用户，请在命令前添加 sudo 。
- `systemctl stop taosd` 指令在执行后并不会马上停止 TDengine 服务，而是会等待系统中必要的落盘工作正常完成。在数据量很大的情况下，这可能会消耗较长时间。
- 如果系统中不支持 `systemd`，也可以用手动运行 `/usr/local/taos/bin/taosd` 方式启动 TDengine 服务。
:::
## TDengine 命令行 (CLI)
为便于检查 TDengine 的状态，执行数据库 (Database) 的各种即席(Ad Hoc)查询，TDengine 提供一命令行应用程序(以下简称为 TDengine CLI) taos。要进入 TDengine 命令行，您只要在安装有 TDengine 的 Linux 终端执行 `taos` 即可。
```bash
taos
```
如果连接服务成功，将会打印出欢迎消息和版本信息。如果失败，则会打印错误消息出来（请参考 [FAQ](/train-faq/faq) 来解决终端连接服务端失败的问题）。 TDengine CLI 的提示符号如下：
```cmd
taos>
```
在 TDengine CLI 中，用户可以通过 SQL 命令来创建/删除数据库、表等，并进行数据库(database)插入查询操作。在终端中运行的 SQL 语句需要以分号结束来运行。示例：
```sql
create database demo;
use demo;
create table t (ts timestamp, speed int);
insert into t values ('2019-07-15 00:00:00', 10);
insert into t values ('2019-07-15 01:00:00', 20);
select * from t;
ts | speed |
========================================
2019-07-15 00:00:00.000 | 10 |
2019-07-15 01:00:00.000 | 20 |
Query OK, 2 row(s) in set (0.003128s)
```
除执行 SQL 语句外，系统管理员还可以从 TDengine CLI 进行检查系统运行状态、添加删除用户账号等操作。TDengine CLI 连同应用驱动也可以独立安装在 Linux 或 Windows 机器上运行，更多细节请参考 [这里](../reference/taos-shell/)
## 使用 taosBenchmark 体验写入速度
启动 TDengine 的服务，在 Linux 终端执行 `taosBenchmark` （曾命名为 `taosdemo`）：
```bash
taosBenchmark
```
该命令将在数据库 test 下面自动创建一张超级表 meters，该超级表下有 1 万张表，表名为 "d0" 到 "d9999"，每张表有 1 万条记录，每条记录有 (ts, current, voltage, phase) 四个字段，时间戳从 "2017-07-14 10:40:00 000" 到 "2017-07-14 10:40:09 999"，每张表带有标签 location 和 groupId，groupId 被设置为 1 到 10， location 被设置为 "California.SanFrancisco" 或者 "California.LosAngeles"。
这条命令很快完成 1 亿条记录的插入。具体时间取决于硬件性能，即使在一台普通的 PC 服务器往往也仅需十几秒。
taosBenchmark 命令本身带有很多选项，配置表的数目、记录条数等等，您可以设置不同参数进行体验，请执行 `taosBenchmark --help` 详细列出。taosBenchmark 详细使用方法请参照 [如何使用 taosBenchmark 对 TDengine 进行性能测试](https://www.taosdata.com/2021/10/09/3111.html)。
## 使用 TDengine CLI 体验查询速度
使用上述 taosBenchmark 插入数据后，可以在 TDengine CLI 输入查询命令，体验查询速度。
查询超级表下记录总条数：
```sql
taos> select count(*) from test.meters;
```
查询 1 亿条记录的平均值、最大值、最小值等：
```sql
taos> select avg(current), max(voltage), min(phase) from test.meters;
```
查询 location="California.SanFrancisco" 的记录总条数：
```sql
taos> select count(*) from test.meters where location="California.SanFrancisco";
```
查询 groupId=10 的所有记录的平均值、最大值、最小值等：
```sql
taos> select avg(current), max(voltage), min(phase) from test.meters where groupId=10;
```
对表 d10 按 10s 进行平均值、最大值和最小值聚合统计：
```sql
taos> select avg(current), max(voltage), min(phase) from test.d10 interval(10s);
```

173
docs/zh/05-get-started/index.md
View File

@ -1,173 +1,14 @@
---
title: 立即开始
description: '从 Docker，安装包或使用 apt-get 快速安装 TDengine, 通过命令行程序TDengine CLI和工具 taosdemo 快速体验 TDengine 功能'
description: '快速设置 TDengine 环境并体验其高效写入和查询'
---
import Tabs from "@theme/Tabs";
import TabItem from "@theme/TabItem";
import PkgInstall from "./\_pkg_install.mdx";
import AptGetInstall from "./\_apt_get_install.mdx";
## 安装
本章主要介绍如何利用 Docker 或者安装包快速设置 TDengine 环境并体验其高效写入和查询。
TDengine 完整的软件包包括服务端（taosd）、用于与第三方系统对接并提供 RESTful 接口的 taosAdapter、应用驱动（taosc）、命令行程序 (CLI，taos) 和一些工具软件，目前 2.X 版服务端 taosd 和 taosAdapter 仅在 Linux 系统上安装和运行，后续将支持 Windows、macOS 等系统。应用驱动 taosc 与 TDengine CLI 可以在 Windows 或 Linux 上安装和运行。TDengine 除了提供多种语言的连接器之外，还通过 [taosAdapter](/reference/taosadapter) 提供 [RESTful 接口](/reference/rest-api)。但在 2.4 之前的版本中没有 taosAdapter，RESTful 接口是由 taosd 内置的 HTTP 服务提供的。
```mdx-code-block
import DocCardList from '@theme/DocCardList';
import {useCurrentSidebarCategory} from '@docusaurus/theme-common';
TDengine 支持 X64/ARM64/MIPS64/Alpha64 硬件平台，后续将支持 ARM32、RISC-V 等 CPU 架构。
<Tabs defaultValue="apt-get">
<TabItem value="docker" label="Docker">
如果已经安装了 docker， 只需执行下面的命令。
```shell
docker run -d -p 6030-6049:6030-6049 -p 6030-6049:6030-6049/udp tdengine/tdengine
```
确定该容器已经启动并且在正常运行
```shell
docker ps
```
进入该容器并执行 bash
```shell
docker exec -it <container name> bash
```
然后就可以执行相关的 Linux 命令操作和访问 TDengine
详细操作方法请参照 [通过 Docker 快速体验 TDengine](/train-faq/docker)。
:::info
从 2.4.0.10 开始，除 taosd 以外，Docker 镜像还包含：taos、taosAdapter、taosdump、taosBenchmark、TDinsight 安装脚本和示例代码。启动 Docker 容器时，将同时启动 taosAdapter 和 taosd，实现对 RESTful 的支持。
:::
</TabItem>
<TabItem value="apt-get" label="apt-get">
<AptGetInstall />
</TabItem>
<TabItem value="pkg" label="安装包">
<PkgInstall />
</TabItem>
<TabItem value="src" label="源码">
如果您希望对 TDengine 贡献代码或对内部实现感兴趣，请参考我们的 [TDengine GitHub 主页](https://github.com/taosdata/TDengine) 下载源码构建和安装.
下载其他组件、最新 Beta 版及之前版本的安装包，请点击[这里](https://www.taosdata.com/cn/all-downloads/)。
</TabItem>
</Tabs>
## 启动
安装后，请使用 `systemctl` 命令来启动 TDengine 的服务进程。
```bash
systemctl start taosd
```
检查服务是否正常工作：
```bash
systemctl status taosd
```
如果 TDengine 服务正常工作，那么您可以通过 TDengine 的命令行程序 `taos` 来访问并体验 TDengine。
:::info
- systemctl 命令需要 _root_ 权限来运行，如果您非 _root_ 用户，请在命令前添加 sudo 。
- 为更好的获得产品反馈，改善产品，TDengine 会采集基本的使用信息，但您可以修改系统配置文件 taos.cfg 里的配置参数 telemetryReporting，将其设为 0，就可将其关闭。
- TDengine 采用 FQDN（一般就是 hostname）作为节点的 ID，为保证正常运行，需要给运行 taosd 的服务器配置好 FQDN，在 TDengine CLI 或应用运行的机器配置好 DNS 服务或 hosts 文件，保证 FQDN 能够解析。
- `systemctl stop taosd` 指令在执行后并不会马上停止 TDengine 服务，而是会等待系统中必要的落盘工作正常完成。在数据量很大的情况下，这可能会消耗较长时间。
TDengine 支持在使用 [`systemd`](https://en.wikipedia.org/wiki/Systemd) 做进程服务管理的 Linux 系统上安装，用 `which systemctl` 命令来检测系统中是否存在 `systemd` 包:
```bash
which systemctl
```
如果系统中不支持 `systemd`，也可以用手动运行 `/usr/local/taos/bin/taosd` 方式启动 TDengine 服务。
:::note
## TDengine 命令行 (CLI)
为便于检查 TDengine 的状态，执行数据库 (Database) 的各种即席(Ad Hoc)查询，TDengine 提供一命令行应用程序(以下简称为 TDengine CLI) taos。要进入 TDengine 命令行，您只要在安装有 TDengine 的 Linux 终端执行 `taos` 即可。
```bash
taos
```
如果连接服务成功，将会打印出欢迎消息和版本信息。如果失败，则会打印错误消息出来（请参考 [FAQ](/train-faq/faq) 来解决终端连接服务端失败的问题）。 TDengine CLI 的提示符号如下：
```cmd
taos>
```
在 TDengine CLI 中，用户可以通过 SQL 命令来创建/删除数据库、表等，并进行数据库(database)插入查询操作。在终端中运行的 SQL 语句需要以分号结束来运行。示例：
```sql
create database demo;
use demo;
create table t (ts timestamp, speed int);
insert into t values ('2019-07-15 00:00:00', 10);
insert into t values ('2019-07-15 01:00:00', 20);
select * from t;
ts | speed |
========================================
2019-07-15 00:00:00.000 | 10 |
2019-07-15 01:00:00.000 | 20 |
Query OK, 2 row(s) in set (0.003128s)
```
除执行 SQL 语句外，系统管理员还可以从 TDengine CLI 进行检查系统运行状态、添加删除用户账号等操作。TDengine CLI 连同应用驱动也可以独立安装在 Linux 或 Windows 机器上运行，更多细节请参考 [这里](../reference/taos-shell/)
## 使用 taosBenchmark 体验写入速度
启动 TDengine 的服务，在 Linux 终端执行 `taosBenchmark` （曾命名为 `taosdemo`）：
```bash
taosBenchmark
```
该命令将在数据库 test 下面自动创建一张超级表 meters，该超级表下有 1 万张表，表名为 "d0" 到 "d9999"，每张表有 1 万条记录，每条记录有 (ts, current, voltage, phase) 四个字段，时间戳从 "2017-07-14 10:40:00 000" 到 "2017-07-14 10:40:09 999"，每张表带有标签 location 和 groupId，groupId 被设置为 1 到 10， location 被设置为 "California.SanFrancisco" 或者 "California.LosAngeles"。
这条命令很快完成 1 亿条记录的插入。具体时间取决于硬件性能，即使在一台普通的 PC 服务器往往也仅需十几秒。
taosBenchmark 命令本身带有很多选项，配置表的数目、记录条数等等，您可以设置不同参数进行体验，请执行 `taosBenchmark --help` 详细列出。taosBenchmark 详细使用方法请参照 [如何使用 taosBenchmark 对 TDengine 进行性能测试](https://www.taosdata.com/2021/10/09/3111.html)。
## 使用 TDengine CLI 体验查询速度
使用上述 taosBenchmark 插入数据后，可以在 TDengine CLI 输入查询命令，体验查询速度。
查询超级表下记录总条数：
```sql
taos> select count(*) from test.meters;
```
查询 1 亿条记录的平均值、最大值、最小值等：
```sql
taos> select avg(current), max(voltage), min(phase) from test.meters;
```
查询 location="California.SanFrancisco" 的记录总条数：
```sql
taos> select count(*) from test.meters where location="California.SanFrancisco";
```
查询 groupId=10 的所有记录的平均值、最大值、最小值等：
```sql
taos> select avg(current), max(voltage), min(phase) from test.meters where groupId=10;
```
对表 d10 按 10s 进行平均值、最大值和最小值聚合统计：
```sql
taos> select avg(current), max(voltage), min(phase) from test.d10 interval(10s);
```
<DocCardList items={useCurrentSidebarCategory().items}/>
```

105
docs/zh/10-cluster/02-cluster-mgmt.md
View File

@ -1,105 +0,0 @@
---
title: 数据节点管理
---
上面已经介绍如何从零开始搭建集群。集群组建完成后，可以随时查看集群中当前的数据节点的状态，还可以添加新的数据节点进行扩容，删除数据节点，甚至手动进行数据节点之间的负载均衡操作。
:::note
以下所有执行命令的操作需要先登陆进 TDengine 系统，必要时请使用 root 权限。
:::
## 查看数据节点
启动 TDengine CLI 程序 taos，然后执行：
```sql
SHOW DNODES;
```
它将列出集群中所有的 dnode，每个 dnode 的 ID，end_point（fqdn:port），状态（ready，offline 等），vnode 数目，还未使用的 vnode 数目等信息。在添加或删除一个数据节点后，可以使用该命令查看。
输出如下（具体内容仅供参考，取决于实际的集群配置）
```
taos> show dnodes;
id | endpoint | vnodes | support_vnodes | status | create_time | note |
============================================================================================================================================
1 | trd01:6030 | 100 | 1024 | ready | 2022-07-15 16:47:47.726 | |
Query OK, 1 rows affected (0.006684s)
```
## 查看虚拟节点组
为充分利用多核技术，并提供横向扩展能力，数据需要分片处理。因此 TDengine 会将一个 DB 的数据切分成多份，存放在多个 vnode 里。这些 vnode 可能分布在多个数据节点 dnode 里，这样就实现了水平扩展。一个 vnode 仅仅属于一个 DB，但一个 DB 可以有多个 vnode。vnode 所在的数据节点是 mnode 根据当前系统资源的情况，自动进行分配的，无需任何人工干预。
启动 CLI 程序 taos，然后执行：
```sql
USE SOME_DATABASE;
SHOW VGROUPS;
```
输出如下（具体内容仅供参考，取决于实际的集群配置）
```
taos> use db;
Database changed.
taos> show vgroups;
vgroup_id | db_name | tables | v1_dnode | v1_status | v2_dnode | v2_status | v3_dnode | v3_status | status | nfiles | file_size | tsma |
================================================================================================================================================================================================
2 | db | 0 | 1 | leader | NULL | NULL | NULL | NULL | NULL | NULL | NULL | 0 |
3 | db | 0 | 1 | leader | NULL | NULL | NULL | NULL | NULL | NULL | NULL | 0 |
4 | db | 0 | 1 | leader | NULL | NULL | NULL | NULL | NULL | NULL | NULL | 0 |
Query OK, 8 row(s) in set (0.001154s)
```
## 添加数据节点
启动 CLI 程序 taos，然后执行：
```sql
CREATE DNODE "fqdn:port";
```
将新数据节点的 End Point 添加进集群的 EP 列表。“fqdn:port“需要用双引号引起来，否则出错。一个数据节点对外服务的 fqdn 和 port 可以通过配置文件 taos.cfg 进行配置，缺省是自动获取。【强烈不建议用自动获取方式来配置 FQDN，可能导致生成的数据节点的 End Point 不是所期望的】
然后启动新加入的数据节点的 taosd 进程，再通过 taos 查看数据节点状态：
```
taos> show dnodes;
id | endpoint | vnodes | support_vnodes | status | create_time | note |
============================================================================================================================================
1 | localhost:6030 | 100 | 1024 | ready | 2022-07-15 16:47:47.726 | |
2 | localhost:7030 | 0 | 1024 | ready | 2022-07-15 16:56:13.670 | |
Query OK, 2 rows affected (0.007031s)
```
从中可以看到两个 dnode 状态都为 ready
## 删除数据节点
先停止要删除的数据节点的 taosd 进程，然后启动 CLI 程序 taos，执行：
```sql
DROP DNODE "fqdn:port";
```
或者
```sql
DROP DNODE dnodeId;
```
通过 “fqdn:port” 或 dnodeID 来指定一个具体的节点都是可以的。其中 fqdn 是被删除的节点的 FQDN，port 是其对外服务器的端口号；dnodeID 可以通过 SHOW DNODES 获得。
:::warning
数据节点一旦被 drop 之后，不能重新加入集群。需要将此节点重新部署（清空数据文件夹）。集群在完成 `drop dnode` 操作之前，会将该 dnode 的数据迁移走。
请注意 `drop dnode` 和 停止 taosd 进程是两个不同的概念，不要混淆：因为删除 dnode 之前要执行迁移数据的操作，因此被删除的 dnode 必须保持在线状态。待删除操作结束之后，才能停止 taosd 进程。
一个数据节点被 drop 之后，其他节点都会感知到这个 dnodeID 的删除操作，任何集群中的节点都不会再接收此 dnodeID 的请求。
dnodeID 是集群自动分配的，不得人工指定。它在生成时是递增的，不会重复。
:::

1
docs/zh/10-cluster/_category_.yml
View File

@ -1 +0,0 @@
label: 集群管理

89
docs/zh/10-cluster/01-deploy.md → docs/zh/10-deployment/01-deploy.md
View File

@ -1,5 +1,6 @@
---
title: 集群部署
sidebar_label: 手动部署
title: 集群部署和管理
---
## 准备工作
@ -54,6 +55,8 @@ fqdn h1.taosdata.com
// 配置本数据节点的端口号，缺省是 6030
serverPort 6030
```
一定要修改的参数是 firstEp 和 fqdn。在每个数据节点，firstEp 需全部配置成一样，但 fqdn 一定要配置成其所在数据节点的值。其他参数可不做任何修改，除非你很清楚为什么要修改。
加入到集群中的数据节点 dnode，下表中涉及集群相关的参数必须完全相同，否则不能成功加入到集群中。
@ -67,8 +70,6 @@ serverPort 6030
## 启动集群
### 启动第一个数据节点
按照《立即开始》里的步骤，启动第一个数据节点，例如 h1.taosdata.com，然后执行 taos，启动 taos shell，从 shell 里执行命令“SHOW DNODES”，如下所示：
```
@ -78,20 +79,17 @@ Copyright (c) 2022 by TAOS Data, Inc. All rights reserved.
Server is Enterprise trial Edition, ver:3.0.0.0 and will never expire.
taos> show dnodes;
id | endpoint | vnodes | support_vnodes | status | create_time | note |
id | endpoint | vnodes | support_vnodes | status | create_time | note |
============================================================================================================================================
1 | h1.taosdata.com:6030 | 0 | 1024 | ready | 2022-07-16 10:50:42.673 | |
1 | h1.taosdata.com:6030 | 0 | 1024 | ready | 2022-07-16 10:50:42.673 | |
Query OK, 1 rows affected (0.007984s)
taos>
taos>
````
```
上述命令里，可以看到刚启动的数据节点的 End Point 是：h1.taos.com:6030，就是这个新集群的 firstEp。
### 启动后续数据节点
## 添加数据节点
将后续的数据节点添加到现有集群，具体有以下几步：
@ -125,3 +123,74 @@ firstEp 这个参数仅仅在该数据节点首次加入集群时有作用，加
两个没有配置 firstEp 参数的数据节点 dnode 启动后，会独立运行起来。这个时候，无法将其中一个数据节点加入到另外一个数据节点，形成集群。无法将两个独立的集群合并成为新的集群。
:::
## 查看数据节点
启动 TDengine CLI 程序 taos，然后执行：
```sql
SHOW DNODES;
```
它将列出集群中所有的 dnode，每个 dnode 的 ID，end_point（fqdn:port），状态（ready，offline 等），vnode 数目，还未使用的 vnode 数目等信息。在添加或删除一个数据节点后，可以使用该命令查看。
输出如下（具体内容仅供参考，取决于实际的集群配置）
```
taos> show dnodes;
id | endpoint | vnodes | support_vnodes | status | create_time | note |
============================================================================================================================================
1 | trd01:6030 | 100 | 1024 | ready | 2022-07-15 16:47:47.726 | |
Query OK, 1 rows affected (0.006684s)
```
## 查看虚拟节点组
为充分利用多核技术，并提供横向扩展能力，数据需要分片处理。因此 TDengine 会将一个 DB 的数据切分成多份，存放在多个 vnode 里。这些 vnode 可能分布在多个数据节点 dnode 里，这样就实现了水平扩展。一个 vnode 仅仅属于一个 DB，但一个 DB 可以有多个 vnode。vnode 所在的数据节点是 mnode 根据当前系统资源的情况，自动进行分配的，无需任何人工干预。
启动 CLI 程序 taos，然后执行：
```sql
USE SOME_DATABASE;
SHOW VGROUPS;
```
输出如下（具体内容仅供参考，取决于实际的集群配置）
```
taos> use db;
Database changed.
taos> show vgroups;
vgroup_id | db_name | tables | v1_dnode | v1_status | v2_dnode | v2_status | v3_dnode | v3_status | status | nfiles | file_size | tsma |
================================================================================================================================================================================================
2 | db | 0 | 1 | leader | NULL | NULL | NULL | NULL | NULL | NULL | NULL | 0 |
3 | db | 0 | 1 | leader | NULL | NULL | NULL | NULL | NULL | NULL | NULL | 0 |
4 | db | 0 | 1 | leader | NULL | NULL | NULL | NULL | NULL | NULL | NULL | 0 |
Query OK, 8 row(s) in set (0.001154s)
```
## 删除数据节点
先停止要删除的数据节点的 taosd 进程，然后启动 CLI 程序 taos，执行：
```sql
DROP DNODE "fqdn:port";
```
或者
```sql
DROP DNODE dnodeId;
```
通过 “fqdn:port” 或 dnodeID 来指定一个具体的节点都是可以的。其中 fqdn 是被删除的节点的 FQDN，port 是其对外服务器的端口号；dnodeID 可以通过 SHOW DNODES 获得。
:::warning
数据节点一旦被 drop 之后，不能重新加入集群。需要将此节点重新部署（清空数据文件夹）。集群在完成 `drop dnode` 操作之前，会将该 dnode 的数据迁移走。
请注意 `drop dnode` 和 停止 taosd 进程是两个不同的概念，不要混淆：因为删除 dnode 之前要执行迁移数据的操作，因此被删除的 dnode 必须保持在线状态。待删除操作结束之后，才能停止 taosd 进程。
一个数据节点被 drop 之后，其他节点都会感知到这个 dnodeID 的删除操作，任何集群中的节点都不会再接收此 dnodeID 的请求。
dnodeID 是集群自动分配的，不得人工指定。它在生成时是递增的，不会重复。
:::

452
docs/zh/10-deployment/03-k8s.md Normal file
View File

@ -0,0 +1,452 @@
---
sidebar_label: Kubernetes
title: 在 Kubernetes 上部署 TDengine 集群
---
## 配置 ConfigMap
为 TDengine 创建 `taoscfg.yaml`，此文件中的配置将作为环境变量传入 TDengine 镜像，更新此配置将导致所有 TDengine POD 重启。
```yaml
---
apiVersion: v1
kind: ConfigMap
metadata:
name: taoscfg
labels:
app: tdengine
data:
CLUSTER: "1"
TAOS_KEEP: "3650"
TAOS_DEBUG_FLAG: "135"
```
## 配置服务
创建一个 service 配置文件：`taosd-service.yaml`，服务名称 `metadata.name` (此处为 "taosd") 将在下一步中使用到。添加 TDengine 所用到的所有端口：
```yaml
---
apiVersion: v1
kind: Service
metadata:
name: "taosd"
labels:
app: "tdengine"
spec:
ports:
- name: tcp6030
protocol: "TCP"
port: 6030
- name: tcp6035
protocol: "TCP"
port: 6035
- name: tcp6041
protocol: "TCP"
port: 6041
- name: udp6030
protocol: "UDP"
port: 6030
- name: udp6031
protocol: "UDP"
port: 6031
- name: udp6032
protocol: "UDP"
port: 6032
- name: udp6033
protocol: "UDP"
port: 6033
- name: udp6034
protocol: "UDP"
port: 6034
- name: udp6035
protocol: "UDP"
port: 6035
- name: udp6036
protocol: "UDP"
port: 6036
- name: udp6037
protocol: "UDP"
port: 6037
- name: udp6038
protocol: "UDP"
port: 6038
- name: udp6039
protocol: "UDP"
port: 6039
- name: udp6040
protocol: "UDP"
port: 6040
selector:
app: "tdengine"
```
## 有状态服务 StatefulSet
根据 Kubernetes 对各类部署的说明，我们将使用 StatefulSet 作为 TDengine 的服务类型，创建文件 `tdengine.yaml`：
```yaml
---
apiVersion: apps/v1
kind: StatefulSet
metadata:
name: "tdengine"
labels:
app: "tdengine"
spec:
serviceName: "taosd"
replicas: 2
updateStrategy:
type: RollingUpdate
selector:
matchLabels:
app: "tdengine"
template:
metadata:
name: "tdengine"
labels:
app: "tdengine"
spec:
containers:
- name: "tdengine"
image: "zitsen/taosd:develop"
imagePullPolicy: "Always"
envFrom:
- configMapRef:
name: taoscfg
ports:
- name: tcp6030
protocol: "TCP"
containerPort: 6030
- name: tcp6035
protocol: "TCP"
containerPort: 6035
- name: tcp6041
protocol: "TCP"
containerPort: 6041
- name: udp6030
protocol: "UDP"
containerPort: 6030
- name: udp6031
protocol: "UDP"
containerPort: 6031
- name: udp6032
protocol: "UDP"
containerPort: 6032
- name: udp6033
protocol: "UDP"
containerPort: 6033
- name: udp6034
protocol: "UDP"
containerPort: 6034
- name: udp6035
protocol: "UDP"
containerPort: 6035
- name: udp6036
protocol: "UDP"
containerPort: 6036
- name: udp6037
protocol: "UDP"
containerPort: 6037
- name: udp6038
protocol: "UDP"
containerPort: 6038
- name: udp6039
protocol: "UDP"
containerPort: 6039
- name: udp6040
protocol: "UDP"
containerPort: 6040
env:
# POD_NAME for FQDN config
- name: POD_NAME
valueFrom:
fieldRef:
fieldPath: metadata.name
# SERVICE_NAME and NAMESPACE for fqdn resolve
- name: SERVICE_NAME
value: "taosd"
- name: STS_NAME
value: "tdengine"
- name: STS_NAMESPACE
valueFrom:
fieldRef:
fieldPath: metadata.namespace
# TZ for timezone settings, we recommend to always set it.
- name: TZ
value: "Asia/Shanghai"
# TAOS_ prefix will configured in taos.cfg, strip prefix and camelCase.
- name: TAOS_SERVER_PORT
value: "6030"
# Must set if you want a cluster.
- name: TAOS_FIRST_EP
value: "$(STS_NAME)-0.$(SERVICE_NAME).$(STS_NAMESPACE).svc.cluster.local:$(TAOS_SERVER_PORT)"
# TAOS_FQND should always be setted in k8s env.
- name: TAOS_FQDN
value: "$(POD_NAME).$(SERVICE_NAME).$(STS_NAMESPACE).svc.cluster.local"
volumeMounts:
- name: taosdata
mountPath: /var/lib/taos
readinessProbe:
exec:
command:
- taos
- -s
- "show mnodes"
initialDelaySeconds: 5
timeoutSeconds: 5000
livenessProbe:
tcpSocket:
port: 6030
initialDelaySeconds: 15
periodSeconds: 20
volumeClaimTemplates:
- metadata:
name: taosdata
spec:
accessModes:
- "ReadWriteOnce"
storageClassName: "csi-rbd-sc"
resources:
requests:
storage: "10Gi"
```
## 启动集群
将前述三个文件添加到 Kubernetes 集群中：
```bash
kubectl apply -f taoscfg.yaml
kubectl apply -f taosd-service.yaml
kubectl apply -f tdengine.yaml
```
上面的配置将生成一个两节点的 TDengine 集群，dnode 是自动配置的，可以使用 `show dnodes` 命令查看当前集群的节点：
```bash
kubectl exec -i -t tdengine-0 -- taos -s "show dnodes"
kubectl exec -i -t tdengine-1 -- taos -s "show dnodes"
```
输出如下：
```
Welcome to the TDengine shell from Linux, Client Version:2.1.1.0
Copyright (c) 2020 by TAOS Data, Inc. All rights reserved.
taos> show dnodes
id | end_point | vnodes | cores | status | role | create_time | offline reason |
======================================================================================================================================
1 | tdengine-0.taosd.default.sv... | 1 | 40 | ready | any | 2021-06-01 17:13:24.181 | |
2 | tdengine-1.taosd.default.sv... | 0 | 40 | ready | any | 2021-06-01 17:14:09.257 | |
Query OK, 2 row(s) in set (0.000997s)
```
## 集群扩容
TDengine 集群支持自动扩容：
```bash
kubectl scale statefulsets tdengine --replicas=4
```
上面命令行中参数 `--replica=4` 表示要将 TDengine 集群扩容到 4 个节点，执行后首先检查 POD 的状态：
```bash
kubectl get pods -l app=tdengine
```
输出如下：
```
NAME READY STATUS RESTARTS AGE
tdengine-0 1/1 Running 0 161m
tdengine-1 1/1 Running 0 161m
tdengine-2 1/1 Running 0 32m
tdengine-3 1/1 Running 0 32m
```
此时 POD 的状态仍然是 Running，TDengine 集群中的 dnode 状态要等 POD 状态为 `ready` 之后才能看到：
```bash
kubectl exec -i -t tdengine-0 -- taos -s "show dnodes"
```
扩容后的四节点 TDengine 集群的 dnode 列表:
```
Welcome to the TDengine shell from Linux, Client Version:2.1.1.0
Copyright (c) 2020 by TAOS Data, Inc. All rights reserved.
taos> show dnodes
id | end_point | vnodes | cores | status | role | create_time | offline reason |
======================================================================================================================================
1 | tdengine-0.taosd.default.sv... | 0 | 40 | ready | any | 2021-06-01 11:58:12.915 | |
2 | tdengine-1.taosd.default.sv... | 0 | 40 | ready | any | 2021-06-01 11:58:33.127 | |
3 | tdengine-2.taosd.default.sv... | 0 | 40 | ready | any | 2021-06-01 14:07:27.078 | |
4 | tdengine-3.taosd.default.sv... | 1 | 40 | ready | any | 2021-06-01 14:07:48.362 | |
Query OK, 4 row(s) in set (0.001293s)
```
## 集群缩容
TDengine 的缩容并没有自动化，我们尝试将一个三节点集群缩容到两节点。
首先，确认一个三节点 TDengine 集群正常工作，在 TDengine CLI 中查看 dnode 的状态：
```bash
taos> show dnodes
id | end_point | vnodes | cores | status | role | create_time | offline reason |
======================================================================================================================================
1 | tdengine-0.taosd.default.sv... | 1 | 40 | ready | any | 2021-06-01 16:27:24.852 | |
2 | tdengine-1.taosd.default.sv... | 0 | 40 | ready | any | 2021-06-01 16:27:53.339 | |
3 | tdengine-2.taosd.default.sv... | 0 | 40 | ready | any | 2021-06-01 16:28:49.787 | |
Query OK, 3 row(s) in set (0.001101s)
```
想要安全的缩容，首先需要将节点从 dnode 列表中移除，也即从集群中移除：
```bash
kubectl exec -i -t tdengine-0 -- taos -s "drop dnode 'tdengine-2.taosd.default.svc.cluster.local:6030'"
```
通过 `show dondes` 命令确认移除成功后，移除相应的 POD：
```bash
kubectl scale statefulsets tdengine --replicas=2
```
最后一个 POD 会被删除，使用 `kubectl get pods -l app=tdengine` 查看集群状态：
```
NAME READY STATUS RESTARTS AGE
tdengine-0 1/1 Running 0 3h40m
tdengine-1 1/1 Running 0 3h40m
```
POD 删除后，需要手动删除 PVC，否则下次扩容时会继续使用以前的数据导致无法正常加入集群。
```bash
kubectl delete pvc taosdata-tdengine-2
```
此时的集群状态是安全的，需要时还可以再次进行扩容：
```bash
kubectl scale statefulsets tdengine --replicas=3
```
`show dnodes` 输出如下：
```
taos> show dnodes
id | end_point | vnodes | cores | status | role | create_time | offline reason |
======================================================================================================================================
1 | tdengine-0.taosd.default.sv... | 1 | 40 | ready | any | 2021-06-01 16:27:24.852 | |
2 | tdengine-1.taosd.default.sv... | 0 | 40 | ready | any | 2021-06-01 16:27:53.339 | |
4 | tdengine-2.taosd.default.sv... | 0 | 40 | ready | any | 2021-06-01 16:40:49.177 | |
```
## 删除集群
完整移除 TDengine 集群，需要分别清理 statefulset、svc、configmap、pvc。
```bash
kubectl delete statefulset -l app=tdengine
kubectl delete svc -l app=tdengine
kubectl delete pvc -l app=tdengine
kubectl delete configmap taoscfg
```
## 常见错误
### 错误一
扩容到四节点之后缩容到两节点，删除的 POD 会进入 offline 状态：
```
Welcome to the TDengine shell from Linux, Client Version:2.1.1.0
Copyright (c) 2020 by TAOS Data, Inc. All rights reserved.
taos> show dnodes
id | end_point | vnodes | cores | status | role | create_time | offline reason |
======================================================================================================================================
1 | tdengine-0.taosd.default.sv... | 0 | 40 | ready | any | 2021-06-01 11:58:12.915 | |
2 | tdengine-1.taosd.default.sv... | 0 | 40 | ready | any | 2021-06-01 11:58:33.127 | |
3 | tdengine-2.taosd.default.sv... | 0 | 40 | offline | any | 2021-06-01 14:07:27.078 | status msg timeout |
4 | tdengine-3.taosd.default.sv... | 1 | 40 | offline | any | 2021-06-01 14:07:48.362 | status msg timeout |
Query OK, 4 row(s) in set (0.001236s)
```
但 `drop dnode` 的行为按不会按照预期进行，且下次集群重启后，所有的 dnode 节点将无法启动 dropping 状态无法退出。
### 错误二
TDengine 集群会持有 replica 参数，如果缩容后的节点数小于这个值，集群将无法使用：
创建一个库使用 replica 参数为 2，插入部分数据：
```bash
kubectl exec -i -t tdengine-0 -- \
taos -s \
"create database if not exists test replica 2;
use test;
create table if not exists t1(ts timestamp, n int);
insert into t1 values(now, 1)(now+1s, 2);"
```
缩容到单节点：
```bash
kubectl scale statefulsets tdengine --replicas=1
```
在 taos shell 中的所有数据库操作将无法成功。
```
taos> show dnodes;
id | end_point | vnodes | cores | status | role | create_time | offline reason |
======================================================================================================================================
1 | tdengine-0.taosd.default.sv... | 2 | 40 | ready | any | 2021-06-01 15:55:52.562 | |
2 | tdengine-1.taosd.default.sv... | 1 | 40 | offline | any | 2021-06-01 15:56:07.212 | status msg timeout |
Query OK, 2 row(s) in set (0.000845s)
taos> show dnodes;
id | end_point | vnodes | cores | status | role | create_time | offline reason |
======================================================================================================================================
1 | tdengine-0.taosd.default.sv... | 2 | 40 | ready | any | 2021-06-01 15:55:52.562 | |
2 | tdengine-1.taosd.default.sv... | 1 | 40 | offline | any | 2021-06-01 15:56:07.212 | status msg timeout |
Query OK, 2 row(s) in set (0.000837s)
taos> use test;
Database changed.
taos> insert into t1 values(now, 3);
DB error: Unable to resolve FQDN (0.013874s)
```

434
docs/zh/10-deployment/05-helm.md Normal file
View File

@ -0,0 +1,434 @@
---
sidebar_label: Helm
title: 使用 Helm 部署 TDengine 集群
---
Helm 是 Kubernetes 的包管理器，上一节使用 Kubernets 部署 TDengine 集群的操作已经足够简单，但 Helm 依然可以提供更强大的能力。
## 安装 Helm
```bash
curl -fsSL -o get_helm.sh \
https://raw.githubusercontent.com/helm/helm/master/scripts/get-helm-3
chmod +x get_helm.sh
./get_helm.sh
```
Helm 会使用 kubectl 和 kubeconfig 的配置来操作 Kubernetes，可以参考 Rancher 安装 Kubernetes 的配置来进行设置。
## 安装 TDengine Chart
TDengine Chart 尚未发布到 Helm 仓库，当前可以从 GitHub 直接下载：
```bash
wget https://github.com/taosdata/TDengine-Operator/raw/main/helm/tdengine-0.3.0.tgz
```
获取当前 Kubernetes 的存储类：
```bash
kubectl get storageclass
```
在 minikube 默认为 standard.
之后，使用 helm 命令安装：
```bash
helm install tdengine tdengine-0.3.0.tgz \
--set storage.className=<your storage class name>
```
在 minikube 环境下，可以设置一个较小的容量避免超出磁盘可用空间：
```bash
helm install tdengine tdengine-0.3.0.tgz \
--set storage.className=standard \
--set storage.dataSize=2Gi \
--set storage.logSize=10Mi
```
部署成功后，TDengine Chart 将会输出操作 TDengine 的说明：
```bash
export POD_NAME=$(kubectl get pods --namespace default \
-l "app.kubernetes.io/name=tdengine,app.kubernetes.io/instance=tdengine" \
-o jsonpath="{.items[0].metadata.name}")
kubectl --namespace default exec $POD_NAME -- taos -s "show dnodes; show mnodes"
kubectl --namespace default exec -it $POD_NAME -- taos
```
可以创建一个表进行测试：
```bash
kubectl --namespace default exec $POD_NAME -- \
taos -s "create database test;
use test;
create table t1 (ts timestamp, n int);
insert into t1 values(now, 1)(now + 1s, 2);
select * from t1;"
```
## 配置 Values
TDengine 支持 `values.yaml` 自定义。
通过 helm show values 可以获取 TDengine Chart 支持的全部 values 列表：
```bash
helm show values tdengine-0.3.0.tgz
```
你可以将结果保存为 values.yaml，之后可以修改其中的各项参数，如 replica 数量，存储类名称，容量大小，TDengine 配置等，然后使用如下命令安装 TDengine 集群：
```bash
helm install tdengine tdengine-0.3.0.tgz -f values.yaml
```
全部参数如下：
```yaml
# Default values for tdengine.
# This is a YAML-formatted file.
# Declare variables to be passed into helm templates.
replicaCount: 1
image:
prefix: tdengine/tdengine
#pullPolicy: Always
# Overrides the image tag whose default is the chart appVersion.
#tag: "2.4.0.5"
service:
# ClusterIP is the default service type, use NodeIP only if you know what you are doing.
type: ClusterIP
ports:
# TCP range required
tcp:
[
6030,
6031,
6032,
6033,
6034,
6035,
6036,
6037,
6038,
6039,
6040,
6041,
6042,
6043,
6044,
6045,
6060,
]
# UDP range 6030-6039
udp: [6030, 6031, 6032, 6033, 6034, 6035, 6036, 6037, 6038, 6039]
arbitrator: true
# Set timezone here, not in taoscfg
timezone: "Asia/Shanghai"
resources:
# We usually recommend not to specify default resources and to leave this as a conscious
# choice for the user. This also increases chances charts run on environments with little
# resources, such as Minikube. If you do want to specify resources, uncomment the following
# lines, adjust them as necessary, and remove the curly braces after 'resources:'.
# limits:
# cpu: 100m
# memory: 128Mi
# requests:
# cpu: 100m
# memory: 128Mi
storage:
# Set storageClassName for pvc. K8s use default storage class if not set.
#
className: ""
dataSize: "100Gi"
logSize: "10Gi"
nodeSelectors:
taosd:
# node selectors
clusterDomainSuffix: ""
# Config settings in taos.cfg file.
#
# The helm/k8s support will use environment variables for taos.cfg,
# converting an upper-snake-cased variable like `TAOS_DEBUG_FLAG`,
# to a camelCase taos config variable `debugFlag`.
#
# See the variable list at https://www.taosdata.com/cn/documentation/administrator .
#
# Note:
# 1. firstEp/secondEp: should not be setted here, it's auto generated at scale-up.
# 2. serverPort: should not be setted, we'll use the default 6030 in many places.
# 3. fqdn: will be auto generated in kubenetes, user should not care about it.
# 4. role: currently role is not supported - every node is able to be mnode and vnode.
#
# Btw, keep quotes "" around the value like below, even the value will be number or not.
taoscfg:
# number of replications, for cluster only
TAOS_REPLICA: "1"
# number of management nodes in the system
TAOS_NUM_OF_MNODES: "1"
# number of days per DB file
# TAOS_DAYS: "10"
# number of days to keep DB file, default is 10 years.
#TAOS_KEEP: "3650"
# cache block size (Mbyte)
#TAOS_CACHE: "16"
# number of cache blocks per vnode
#TAOS_BLOCKS: "6"
# minimum rows of records in file block
#TAOS_MIN_ROWS: "100"
# maximum rows of records in file block
#TAOS_MAX_ROWS: "4096"
#
# TAOS_NUM_OF_THREADS_PER_CORE: number of threads per CPU core
#TAOS_NUM_OF_THREADS_PER_CORE: "1.0"
#
# TAOS_NUM_OF_COMMIT_THREADS: number of threads to commit cache data
#TAOS_NUM_OF_COMMIT_THREADS: "4"
#
# TAOS_RATIO_OF_QUERY_CORES:
# the proportion of total CPU cores available for query processing
# 2.0: the query threads will be set to double of the CPU cores.
# 1.0: all CPU cores are available for query processing [default].
# 0.5: only half of the CPU cores are available for query.
# 0.0: only one core available.
#TAOS_RATIO_OF_QUERY_CORES: "1.0"
#
# TAOS_KEEP_COLUMN_NAME:
# the last_row/first/last aggregator will not change the original column name in the result fields
#TAOS_KEEP_COLUMN_NAME: "0"
# enable/disable backuping vnode directory when removing vnode
#TAOS_VNODE_BAK: "1"
# enable/disable installation / usage report
#TAOS_TELEMETRY_REPORTING: "1"
# enable/disable load balancing
#TAOS_BALANCE: "1"
# max timer control blocks
#TAOS_MAX_TMR_CTRL: "512"
# time interval of system monitor, seconds
#TAOS_MONITOR_INTERVAL: "30"
# number of seconds allowed for a dnode to be offline, for cluster only
#TAOS_OFFLINE_THRESHOLD: "8640000"
# RPC re-try timer, millisecond
#TAOS_RPC_TIMER: "1000"
# RPC maximum time for ack, seconds.
#TAOS_RPC_MAX_TIME: "600"
# time interval of dnode status reporting to mnode, seconds, for cluster only
#TAOS_STATUS_INTERVAL: "1"
# time interval of heart beat from shell to dnode, seconds
#TAOS_SHELL_ACTIVITY_TIMER: "3"
# minimum sliding window time, milli-second
#TAOS_MIN_SLIDING_TIME: "10"
# minimum time window, milli-second
#TAOS_MIN_INTERVAL_TIME: "10"
# maximum delay before launching a stream computation, milli-second
#TAOS_MAX_STREAM_COMP_DELAY: "20000"
# maximum delay before launching a stream computation for the first time, milli-second
#TAOS_MAX_FIRST_STREAM_COMP_DELAY: "10000"
# retry delay when a stream computation fails, milli-second
#TAOS_RETRY_STREAM_COMP_DELAY: "10"
# the delayed time for launching a stream computation, from 0.1(default, 10% of whole computing time window) to 0.9
#TAOS_STREAM_COMP_DELAY_RATIO: "0.1"
# max number of vgroups per db, 0 means configured automatically
#TAOS_MAX_VGROUPS_PER_DB: "0"
# max number of tables per vnode
#TAOS_MAX_TABLES_PER_VNODE: "1000000"
# the number of acknowledgments required for successful data writing
#TAOS_QUORUM: "1"
# enable/disable compression
#TAOS_COMP: "2"
# write ahead log (WAL) level, 0: no wal; 1: write wal, but no fysnc; 2: write wal, and call fsync
#TAOS_WAL_LEVEL: "1"
# if walLevel is set to 2, the cycle of fsync being executed, if set to 0, fsync is called right away
#TAOS_FSYNC: "3000"
# the compressed rpc message, option:
# -1 (no compression)
# 0 (all message compressed),
# > 0 (rpc message body which larger than this value will be compressed)
#TAOS_COMPRESS_MSG_SIZE: "-1"
# max length of an SQL
#TAOS_MAX_SQL_LENGTH: "1048576"
# the maximum number of records allowed for super table time sorting
#TAOS_MAX_NUM_OF_ORDERED_RES: "100000"
# max number of connections allowed in dnode
#TAOS_MAX_SHELL_CONNS: "5000"
# max number of connections allowed in client
#TAOS_MAX_CONNECTIONS: "5000"
# stop writing logs when the disk size of the log folder is less than this value
#TAOS_MINIMAL_LOG_DIR_G_B: "0.1"
# stop writing temporary files when the disk size of the tmp folder is less than this value
#TAOS_MINIMAL_TMP_DIR_G_B: "0.1"
# if disk free space is less than this value, taosd service exit directly within startup process
#TAOS_MINIMAL_DATA_DIR_G_B: "0.1"
# One mnode is equal to the number of vnode consumed
#TAOS_MNODE_EQUAL_VNODE_NUM: "4"
# enbale/disable http service
#TAOS_HTTP: "1"
# enable/disable system monitor
#TAOS_MONITOR: "1"
# enable/disable recording the SQL statements via restful interface
#TAOS_HTTP_ENABLE_RECORD_SQL: "0"
# number of threads used to process http requests
#TAOS_HTTP_MAX_THREADS: "2"
# maximum number of rows returned by the restful interface
#TAOS_RESTFUL_ROW_LIMIT: "10240"
# The following parameter is used to limit the maximum number of lines in log files.
# max number of lines per log filters
# numOfLogLines 10000000
# enable/disable async log
#TAOS_ASYNC_LOG: "0"
#
# time of keeping log files, days
#TAOS_LOG_KEEP_DAYS: "0"
# The following parameters are used for debug purpose only.
# debugFlag 8 bits mask: FILE-SCREEN-UNUSED-HeartBeat-DUMP-TRACE_WARN-ERROR
# 131: output warning and error
# 135: output debug, warning and error
# 143: output trace, debug, warning and error to log
# 199: output debug, warning and error to both screen and file
# 207: output trace, debug, warning and error to both screen and file
#
# debug flag for all log type, take effect when non-zero value\
#TAOS_DEBUG_FLAG: "143"
# enable/disable recording the SQL in taos client
#TAOS_ENABLE_RECORD_SQL: "0"
# generate core file when service crash
#TAOS_ENABLE_CORE_FILE: "1"
# maximum display width of binary and nchar fields in the shell. The parts exceeding this limit will be hidden
#TAOS_MAX_BINARY_DISPLAY_WIDTH: "30"
# enable/disable stream (continuous query)
#TAOS_STREAM: "1"
# in retrieve blocking model, only in 50% query threads will be used in query processing in dnode
#TAOS_RETRIEVE_BLOCKING_MODEL: "0"
# the maximum allowed query buffer size in MB during query processing for each data node
# -1 no limit (default)
# 0 no query allowed, queries are disabled
#TAOS_QUERY_BUFFER_SIZE: "-1"
```
## 扩容
关于扩容可参考上一节的说明，有一些额外的操作需要从 helm 的部署中获取。
首先，从部署中获取 StatefulSet 的名称。
```bash
export STS_NAME=$(kubectl get statefulset \
-l "app.kubernetes.io/name=tdengine" \
-o jsonpath="{.items[0].metadata.name}")
```
扩容操作极其简单，增加 replica 即可。以下命令将 TDengine 扩充到三节点：
```bash
kubectl scale --replicas 3 statefulset/$STS_NAME
```
使用命令 `show dnodes` 和 `show mnodes` 检查是否扩容成功。
## 缩容
:::warning
缩容操作并没有完整测试，可能造成数据风险，请谨慎使用。
:::
获取需要缩容的 dnode 列表，并手动 Drop。
```bash
kubectl --namespace default exec $POD_NAME -- \
cat /var/lib/taos/dnode/dnodeEps.json \
| jq '.dnodeInfos[1:] |map(.dnodeFqdn + ":" + (.dnodePort|tostring)) | .[]' -r
kubectl --namespace default exec $POD_NAME -- taos -s "show dnodes"
kubectl --namespace default exec $POD_NAME -- taos -s 'drop dnode "<you dnode in list>"'
```
## 删除集群
Helm 管理下，清理操作也变得简单：
```bash
helm uninstall tdengine
```
但 Helm 也不会自动移除 PVC，需要手动获取 PVC 然后删除掉。

1
docs/zh/10-deployment/_category_.yml Normal file
View File

@ -0,0 +1 @@
label: 部署集群

4
docs/zh/10-cluster/index.md → docs/zh/10-deployment/index.md
View File

@ -1,10 +1,10 @@
---
title: 集群管理
title: 部署集群
---
TDengine 支持集群，提供水平扩展的能力。如果需要获得更高的处理能力，只需要多增加节点即可。TDengine 采用虚拟节点技术，将一个节点虚拟化为多个虚拟节点，以实现负载均衡。同时，TDengine可以将多个节点上的虚拟节点组成虚拟节点组，通过多副本机制，以保证供系统的高可用。TDengine的集群功能完全开源。
本章节主要介绍集群的部署、维护，以及如何实现高可用和负载均衡。
本章节主要介绍如何在主机上人工部署集群，以及如何使用 Kubernetes 和 Helm部署集群。
```mdx-code-block
import DocCardList from '@theme/DocCardList';

222
docs/zh/13-operation/01-pkg-install.md
View File

@ -6,139 +6,50 @@ description: 安装、卸载、启动、停止和升级
import Tabs from "@theme/Tabs";
import TabItem from "@theme/TabItem";
TDengine 开源版本提供 deb 和 rpm 格式安装包，用户可以根据自己的运行环境选择合适的安装包。其中 deb 支持 Debian/Ubuntu 及衍生系统，rpm 支持 CentOS/RHEL/SUSE 及衍生系统。同时我们也为企业用户提供 tar.gz 格式安装包。
本节将介绍一些关于安装和卸载更深层次的内容，以及升级的注意事项。
## 安装
<Tabs>
<TabItem label="Deb 安装" value="debinst">
关于安装，请参考 [使用安装包立即开始](../get-started/package)
1、从官网下载获得 deb 安装包，例如 TDengine-server-2.4.0.7-Linux-x64.deb；
2、进入到 TDengine-server-2.4.0.7-Linux-x64.deb 安装包所在目录，执行如下的安装命令：
## 安装目录说明
TDengine 成功安装后，主安装目录是 /usr/local/taos，目录内容如下：
```
$ sudo dpkg -i TDengine-server-2.4.0.7-Linux-x64.deb
(Reading database ... 137504 files and directories currently installed.)
Preparing to unpack TDengine-server-2.4.0.7-Linux-x64.deb ...
TDengine is removed successfully!
Unpacking tdengine (2.4.0.7) over (2.4.0.7) ...
Setting up tdengine (2.4.0.7) ...
Start to install TDengine...
System hostname is: ubuntu-1804
Enter FQDN:port (like h1.taosdata.com:6030) of an existing TDengine cluster node to join
OR leave it blank to build one:
Enter your email address for priority support or enter empty to skip:
Created symlink /etc/systemd/system/multi-user.target.wants/taosd.service → /etc/systemd/system/taosd.service.
To configure TDengine : edit /etc/taos/taos.cfg
To start TDengine : sudo systemctl start taosd
To access TDengine : taos -h ubuntu-1804 to login into TDengine server
TDengine is installed successfully!
```
</TabItem>
<TabItem label="RPM 安装" value="rpminst">
1、从官网下载获得 rpm 安装包，例如 TDengine-server-2.4.0.7-Linux-x64.rpm；
2、进入到 TDengine-server-2.4.0.7-Linux-x64.rpm 安装包所在目录，执行如下的安装命令：
```
$ sudo rpm -ivh TDengine-server-2.4.0.7-Linux-x64.rpm
Preparing... ################################# [100%]
Updating / installing...
1:tdengine-2.4.0.7-3 ################################# [100%]
Start to install TDengine...
System hostname is: centos7
Enter FQDN:port (like h1.taosdata.com:6030) of an existing TDengine cluster node to join
OR leave it blank to build one:
Enter your email address for priority support or enter empty to skip:
Created symlink from /etc/systemd/system/multi-user.target.wants/taosd.service to /etc/systemd/system/taosd.service.
To configure TDengine : edit /etc/taos/taos.cfg
To start TDengine : sudo systemctl start taosd
To access TDengine : taos -h centos7 to login into TDengine server
TDengine is installed successfully!
```
</TabItem>
<TabItem label="tar.gz 安装" value="tarinst">
1、从官网下载获得 tar.gz 安装包，例如 TDengine-server-2.4.0.7-Linux-x64.tar.gz；
2、进入到 TDengine-server-2.4.0.7-Linux-x64.tar.gz 安装包所在目录，先解压文件后，进入子目录，执行其中的 install.sh 安装脚本：
```
$ tar xvzf TDengine-enterprise-server-2.4.0.7-Linux-x64.tar.gz
TDengine-enterprise-server-2.4.0.7/
TDengine-enterprise-server-2.4.0.7/driver/
TDengine-enterprise-server-2.4.0.7/driver/vercomp.txt
TDengine-enterprise-server-2.4.0.7/driver/libtaos.so.2.4.0.7
TDengine-enterprise-server-2.4.0.7/install.sh
TDengine-enterprise-server-2.4.0.7/examples/
...
$ cd /usr/local/taos
$ ll
total 43816
drwxrwxr-x 3 ubuntu ubuntu 4096 Feb 22 09:31 ./
drwxr-xr-x 20 ubuntu ubuntu 4096 Feb 22 09:30 ../
drwxrwxr-x 4 ubuntu ubuntu 4096 Feb 22 09:30 TDengine-enterprise-server-2.4.0.7/
-rw-rw-r-- 1 ubuntu ubuntu 44852544 Feb 22 09:31 TDengine-enterprise-server-2.4.0.7-Linux-x64.tar.gz
$ cd TDengine-enterprise-server-2.4.0.7/
$ ll
total 40784
drwxrwxr-x 4 ubuntu ubuntu 4096 Feb 22 09:30 ./
drwxrwxr-x 3 ubuntu ubuntu 4096 Feb 22 09:31 ../
drwxrwxr-x 2 ubuntu ubuntu 4096 Feb 22 09:30 driver/
drwxrwxr-x 10 ubuntu ubuntu 4096 Feb 22 09:30 examples/
-rwxrwxr-x 1 ubuntu ubuntu 33294 Feb 22 09:30 install.sh*
-rw-rw-r-- 1 ubuntu ubuntu 41704288 Feb 22 09:30 taos.tar.gz
$ sudo ./install.sh
Start to update TDengine...
Created symlink /etc/systemd/system/multi-user.target.wants/taosd.service → /etc/systemd/system/taosd.service.
Nginx for TDengine is updated successfully!
To configure TDengine : edit /etc/taos/taos.cfg
To configure Taos Adapter (if has) : edit /etc/taos/taosadapter.toml
To start TDengine : sudo systemctl start taosd
To access TDengine : use taos -h ubuntu-1804 in shell OR from http://127.0.0.1:6060
TDengine is updated successfully!
Install taoskeeper as a standalone service
taoskeeper is installed, enable it by `systemctl enable taoskeeper`
$ ll
total 28
drwxr-xr-x 7 root root 4096 Feb 22 09:34 ./
drwxr-xr-x 12 root root 4096 Feb 22 09:34 ../
drwxr-xr-x 2 root root 4096 Feb 22 09:34 bin/
drwxr-xr-x 2 root root 4096 Feb 22 09:34 cfg/
lrwxrwxrwx 1 root root 13 Feb 22 09:34 data -> /var/lib/taos/
drwxr-xr-x 2 root root 4096 Feb 22 09:34 driver/
drwxr-xr-x 10 root root 4096 Feb 22 09:34 examples/
drwxr-xr-x 2 root root 4096 Feb 22 09:34 include/
lrwxrwxrwx 1 root root 13 Feb 22 09:34 log -> /var/log/taos/
```
:::info
install.sh 安装脚本在执行过程中，会通过命令行交互界面询问一些配置信息。如果希望采取无交互安装方式，那么可以用 -e no 参数来执行 install.sh 脚本。运行 `./install.sh -h` 指令可以查看所有参数的详细说明信息。
:::
</TabItem>
</Tabs>
:::note
当安装第一个节点时，出现 Enter FQDN：提示的时候，不需要输入任何内容。只有当安装第二个或以后更多的节点时，才需要输入已有集群中任何一个可用节点的 FQDN，支持该新节点加入集群。当然也可以不输入，而是在新节点启动前，配置到新节点的配置文件中。
:::
- 自动生成配置文件目录、数据库目录、日志目录。
- 配置文件缺省目录：/etc/taos/taos.cfg， 软链接到 /usr/local/taos/cfg/taos.cfg；
- 数据库缺省目录：/var/lib/taos， 软链接到 /usr/local/taos/data；
- 日志缺省目录：/var/log/taos， 软链接到 /usr/local/taos/log；
- /usr/local/taos/bin 目录下的可执行文件，会软链接到 /usr/bin 目录下；
- /usr/local/taos/driver 目录下的动态库文件，会软链接到 /usr/lib 目录下；
- /usr/local/taos/include 目录下的头文件，会软链接到到 /usr/include 目录下；
## 卸载
<Tabs>
<TabItem label="apt-get 卸载" value="aptremove">
内容 TBD
</TabItem>
<TabItem label="Deb 卸载" value="debuninst">
卸载命令如下:
@ -180,88 +91,33 @@ taosKeeper is removed successfully!
</Tabs>
:::info
- TDengine 提供了多种安装包，但最好不要在一个系统上同时使用 tar.gz 安装包和 deb 或 rpm 安装包。否则会相互影响，导致在使用时出现问题。
- 对于 deb 包安装后，如果安装目录被手工误删了部分，出现卸载、或重新安装不能成功。此时，需要清除 TDengine 包的安装信息，执行如下命令：
```
$ sudo rm -f /var/lib/dpkg/info/tdengine*
```
```
$ sudo rm -f /var/lib/dpkg/info/tdengine*
```
然后再重新进行安装就可以了。
- 对于 rpm 包安装后，如果安装目录被手工误删了部分，出现卸载、或重新安装不能成功。此时，需要清除 TDengine 包的安装信息，执行如下命令：
```
$ sudo rpm -e --noscripts tdengine
```
```
$ sudo rpm -e --noscripts tdengine
```
然后再重新进行安装就可以了。
:::
## 安装目录说明
TDengine 成功安装后，主安装目录是 /usr/local/taos，目录内容如下：
```
$ cd /usr/local/taos
$ ll
$ ll
total 28
drwxr-xr-x 7 root root 4096 Feb 22 09:34 ./
drwxr-xr-x 12 root root 4096 Feb 22 09:34 ../
drwxr-xr-x 2 root root 4096 Feb 22 09:34 bin/
drwxr-xr-x 2 root root 4096 Feb 22 09:34 cfg/
lrwxrwxrwx 1 root root 13 Feb 22 09:34 data -> /var/lib/taos/
drwxr-xr-x 2 root root 4096 Feb 22 09:34 driver/
drwxr-xr-x 10 root root 4096 Feb 22 09:34 examples/
drwxr-xr-x 2 root root 4096 Feb 22 09:34 include/
lrwxrwxrwx 1 root root 13 Feb 22 09:34 log -> /var/log/taos/
```
- 自动生成配置文件目录、数据库目录、日志目录。
- 配置文件缺省目录：/etc/taos/taos.cfg， 软链接到 /usr/local/taos/cfg/taos.cfg；
- 数据库缺省目录：/var/lib/taos， 软链接到 /usr/local/taos/data；
- 日志缺省目录：/var/log/taos， 软链接到 /usr/local/taos/log；
- /usr/local/taos/bin 目录下的可执行文件，会软链接到 /usr/bin 目录下；
- /usr/local/taos/driver 目录下的动态库文件，会软链接到 /usr/lib 目录下；
- /usr/local/taos/include 目录下的头文件，会软链接到到 /usr/include 目录下；
## 卸载和更新文件说明
卸载安装包的时候，将保留配置文件、数据库文件和日志文件，即 /etc/taos/taos.cfg 、 /var/lib/taos 、 /var/log/taos 。如果用户确认后不需保留，可以手工删除，但一定要慎重，因为删除后，数据将永久丢失，不可以恢复！
如果是更新安装，当缺省配置文件（ /etc/taos/taos.cfg ）存在时，仍然使用已有的配置文件，安装包中携带的配置文件修改为 taos.cfg.orig 保存在 /usr/local/taos/cfg/ 目录，可以作为设置配置参数的参考样例；如果不存在配置文件，就使用安装包中自带的配置文件。
## 启动和停止
TDengine 使用 Linux 系统的 systemd/systemctl/service 来管理系统的启动和、停止、重启操作。TDengine 的服务进程是 taosd，默认情况下 TDengine 在系统启动后将自动启动。DBA 可以通过 systemd/systemctl/service 手动操作停止、启动、重新启动服务。
以 systemctl 为例，命令如下：
- 启动服务进程：`systemctl start taosd`
- 停止服务进程：`systemctl stop taosd`
- 重启服务进程：`systemctl restart taosd`
- 查看服务状态：`systemctl status taosd`
注意：TDengine 在 2.4 版本之后包含一个独立组件 taosAdapter 需要使用 systemctl 命令管理 taosAdapter 服务的启动和停止。
如果服务进程处于活动状态，则 status 指令会显示如下的相关信息：
```
Active: active (running)
```
如果后台服务进程处于停止状态，则 status 指令会显示如下的相关信息：
```
Active: inactive (dead)
```
## 升级
升级分为两个层面：升级安装包 和 升级运行中的实例。
@ -280,4 +136,4 @@ TDengine 使用 Linux 系统的 systemd/systemctl/service 来管理系统的启
:::warning
TDengine 不保证低版本能够兼容高版本的数据，所以任何时候都不推荐降级
:::
:::

0
docs/zh/10-cluster/03-high-availability.md → docs/zh/21-tdinternal/03-high-availability.md
View File

0
docs/zh/10-cluster/04-load-balance.md → docs/zh/21-tdinternal/05-load-balance.md
View File

1
include/common/tcommon.h
View File

@ -103,6 +103,7 @@ typedef struct SDataBlockInfo {
int16_t hasVarCol;
uint32_t capacity;
// TODO: optimize and remove following
int64_t version; // used for stream, and need serialization
int32_t childId; // used for stream, do not serialize
EStreamType type; // used for stream, do not serialize
STimeWindow calWin; // used for stream, do not serialize

3
include/common/tmsg.h
View File

@ -438,7 +438,7 @@ static FORCE_INLINE int32_t tDecodeSSchemaWrapperEx(SDecoder* pDecoder, SSchemaW
return 0;
}
STSchema* tdGetSTSChemaFromSSChema(SSchema** pSchema, int32_t nCols);
STSchema* tdGetSTSChemaFromSSChema(SSchema* pSchema, int32_t nCols, int32_t sver);
typedef struct {
char name[TSDB_TABLE_FNAME_LEN];
@ -1359,6 +1359,7 @@ typedef struct {
int32_t numOfCols;
int64_t skey;
int64_t ekey;
int64_t version; // for stream
char data[];
} SRetrieveTableRsp;

1
include/common/tname.h
View File

@ -50,6 +50,7 @@ bool tNameIsValid(const SName* name);
const char* tNameGetTableName(const SName* name);
int32_t tNameGetDbName(const SName* name, char* dst);
const char* tNameGetDbNameP(const SName* name);
int32_t tNameGetFullDbName(const SName* name, char* dst);

2
include/libs/executor/executor.h
View File

@ -64,7 +64,7 @@ qTaskInfo_t qCreateStreamExecTaskInfo(void* msg, SReadHandle* readers);
* @param SReadHandle
* @return
*/
qTaskInfo_t qCreateQueueExecTaskInfo(void* msg, SReadHandle* readers, int32_t* numOfCols);
qTaskInfo_t qCreateQueueExecTaskInfo(void* msg, SReadHandle* readers, int32_t* numOfCols, SSchemaWrapper** pSchemaWrapper);
/**
* Set the input data block for the stream scan.

1
include/libs/function/functionMgt.h
View File

@ -197,6 +197,7 @@ bool fmIsSystemInfoFunc(int32_t funcId);
bool fmIsImplicitTsFunc(int32_t funcId);
bool fmIsClientPseudoColumnFunc(int32_t funcId);
bool fmIsMultiRowsFunc(int32_t funcId);
bool fmIsKeepOrderFunc(int32_t funcId);
int32_t fmGetDistMethod(const SFunctionNode* pFunc, SFunctionNode** pPartialFunc, SFunctionNode** pMergeFunc);

11
include/libs/nodes/plannodes.h
View File

@ -29,9 +29,17 @@ extern "C" {
typedef enum EDataOrderLevel {
DATA_ORDER_LEVEL_NONE = 1,
DATA_ORDER_LEVEL_IN_BLOCK,
DATA_ORDER_LEVEL_IN_GROUP
DATA_ORDER_LEVEL_IN_GROUP,
DATA_ORDER_LEVEL_GLOBAL
} EDataOrderLevel;
typedef enum EGroupAction {
GROUP_ACTION_NONE = 1,
GROUP_ACTION_SET,
GROUP_ACTION_KEEP,
GROUP_ACTION_CLEAR
} EGroupAction;
typedef struct SLogicNode {
ENodeType type;
SNodeList* pTargets; // SColumnNode
@ -44,6 +52,7 @@ typedef struct SLogicNode {
SNode* pSlimit;
EDataOrderLevel requireDataOrder; // requirements for input data
EDataOrderLevel resultDataOrder; // properties of the output data
EGroupAction groupAction;
} SLogicNode;
typedef enum EScanType {

6
include/libs/qcom/query.h
View File

@ -193,7 +193,7 @@ int32_t taosAsyncExec(__async_exec_fn_t execFn, void* execParam, int32_t* code);
void destroySendMsgInfo(SMsgSendInfo* pMsgBody);
int32_t asyncSendMsgToServerExt(void* pTransporter, SEpSet* epSet, int64_t* pTransporterId, const SMsgSendInfo* pInfo,
int32_t asyncSendMsgToServerExt(void* pTransporter, SEpSet* epSet, int64_t* pTransporterId, SMsgSendInfo* pInfo,
bool persistHandle, void* ctx);
/**
@ -205,7 +205,7 @@ int32_t asyncSendMsgToServerExt(void* pTransporter, SEpSet* epSet, int64_t* pTra
* @param pInfo
* @return
*/
int32_t asyncSendMsgToServer(void* pTransporter, SEpSet* epSet, int64_t* pTransporterId, const SMsgSendInfo* pInfo);
int32_t asyncSendMsgToServer(void* pTransporter, SEpSet* epSet, int64_t* pTransporterId, SMsgSendInfo* pInfo);
int32_t queryBuildUseDbOutput(SUseDbOutput* pOut, SUseDbRsp* usedbRsp);
@ -260,6 +260,8 @@ extern int32_t (*queryProcessMsgRsp[TDMT_MAX])(void* output, char* msg, int32_t
#define REQUEST_TOTAL_EXEC_TIMES 2
#define IS_SYS_DBNAME(_dbname) (((*(_dbname) == 'i') && (0 == strcmp(_dbname, TSDB_INFORMATION_SCHEMA_DB))) || ((*(_dbname) == 'p') && (0 == strcmp(_dbname, TSDB_PERFORMANCE_SCHEMA_DB))))
#define qFatal(...) \
do { \
if (qDebugFlag & DEBUG_FATAL) { \

1
include/libs/stream/tstream.h
View File

@ -142,6 +142,7 @@ static FORCE_INLINE void* streamQueueNextItem(SStreamQueue* queue) {
ASSERT(queue->qItem != NULL);
return streamQueueCurItem(queue);
} else {
queue->qItem = NULL;
taosGetQitem(queue->qall, &queue->qItem);
if (queue->qItem == NULL) {
taosReadAllQitems(queue->queue, queue->qall);

3
include/libs/stream/tstreamUpdate.h
View File

@ -33,11 +33,12 @@ typedef struct SUpdateInfo {
int64_t watermark;
TSKEY minTS;
SScalableBf* pCloseWinSBF;
SHashObj* pMap;
} SUpdateInfo;
SUpdateInfo *updateInfoInitP(SInterval* pInterval, int64_t watermark);
SUpdateInfo *updateInfoInit(int64_t interval, int32_t precision, int64_t watermark);
bool updateInfoIsUpdated(SUpdateInfo *pInfo, tb_uid_t tableId, TSKEY ts);
bool updateInfoIsUpdated(SUpdateInfo *pInfo, uint64_t tableId, TSKEY ts);
void updateInfoDestroy(SUpdateInfo *pInfo);
void updateInfoAddCloseWindowSBF(SUpdateInfo *pInfo);
void updateInfoDestoryColseWinSBF(SUpdateInfo *pInfo);

18
include/libs/transport/trpc.h
View File

@ -124,18 +124,16 @@ void *rpcReallocCont(void *ptr, int32_t contLen);
// Because taosd supports multi-process mode
// These functions should not be used on the server side
// Please use tmsg<xx> functions, which are defined in tmsgcb.h
void rpcSendRequest(void *thandle, const SEpSet *pEpSet, SRpcMsg *pMsg, int64_t *rid);
void rpcSendResponse(const SRpcMsg *pMsg);
void rpcRegisterBrokenLinkArg(SRpcMsg *msg);
void rpcReleaseHandle(void *handle, int8_t type); // just release conn to rpc instance, no close sock
int rpcSendRequest(void *thandle, const SEpSet *pEpSet, SRpcMsg *pMsg, int64_t *rid);
int rpcSendResponse(const SRpcMsg *pMsg);
int rpcRegisterBrokenLinkArg(SRpcMsg *msg);
int rpcReleaseHandle(void *handle, int8_t type); // just release conn to rpc instance, no close sock
// These functions will not be called in the child process
void rpcSendRedirectRsp(void *pConn, const SEpSet *pEpSet);
void rpcSendRequestWithCtx(void *thandle, const SEpSet *pEpSet, SRpcMsg *pMsg, int64_t *rid, SRpcCtx *ctx);
int32_t rpcGetConnInfo(void *thandle, SRpcConnInfo *pInfo);
void rpcSendRecv(void *shandle, SEpSet *pEpSet, SRpcMsg *pReq, SRpcMsg *pRsp);
void rpcSetDefaultAddr(void *thandle, const char *ip, const char *fqdn);
void* rpcAllocHandle();
int rpcSendRequestWithCtx(void *thandle, const SEpSet *pEpSet, SRpcMsg *pMsg, int64_t *rid, SRpcCtx *ctx);
int rpcSendRecv(void *shandle, SEpSet *pEpSet, SRpcMsg *pReq, SRpcMsg *pRsp);
int rpcSetDefaultAddr(void *thandle, const char *ip, const char *fqdn);
void *rpcAllocHandle();
#ifdef __cplusplus
}

4
include/os/osSignal.h
View File

@ -44,7 +44,11 @@ extern "C" {
#define SIGBREAK 1234
#endif
#ifdef WINDOWS
typedef BOOL (*FSignalHandler)(DWORD fdwCtrlType);
#else
typedef void (*FSignalHandler)(int32_t signum, void *sigInfo, void *context);
#endif
void taosSetSignal(int32_t signum, FSignalHandler sigfp);
void taosIgnSignal(int32_t signum);
void taosDflSignal(int32_t signum);

2
include/os/osSysinfo.h
View File

@ -33,7 +33,7 @@ typedef struct {
SDiskSize size;
} SDiskSpace;
bool taosCheckSystemIsSmallEnd();
bool taosCheckSystemIsLittleEnd();
void taosGetSystemInfo();
int32_t taosGetEmail(char *email, int32_t maxLen);
int32_t taosGetOsReleaseName(char *releaseName, int32_t maxLen);

3
include/os/osSystem.h
View File

@ -29,9 +29,6 @@ extern "C" {
#define tcgetattr TCGETATTR_FUNC_TAOS_FORBID
#endif
#define TAOS_CONSOLE_PROMPT_HEADER "taos> "
#define TAOS_CONSOLE_PROMPT_CONTINUE " -> "
typedef struct TdCmd *TdCmdPtr;
TdCmdPtr taosOpenCmd(const char* cmd);

2
source/client/inc/clientStmt.h
View File

@ -72,7 +72,6 @@ typedef struct SStmtBindInfo {
typedef struct SStmtExecInfo {
int32_t affectedRows;
SRequestObj* pRequest;
SHashObj* pVgHash;
SHashObj* pBlockHash;
bool autoCreateTbl;
} SStmtExecInfo;
@ -88,6 +87,7 @@ typedef struct SStmtSQLInfo {
SArray* nodeList;
SStmtQueryResInfo queryRes;
bool autoCreateTbl;
SHashObj* pVgHash;
} SStmtSQLInfo;
typedef struct STscStmt {

2
source/client/src/clientEnv.c
View File

@ -88,7 +88,7 @@ void closeTransporter(SAppInstInfo *pAppInfo) {
static bool clientRpcRfp(int32_t code, tmsg_t msgType) {
if (NEED_REDIRECT_ERROR(code)) {
if (msgType == TDMT_SCH_QUERY || msgType == TDMT_SCH_MERGE_QUERY || msgType == TDMT_SCH_FETCH ||
msgType == TDMT_SCH_MERGE_FETCH) {
msgType == TDMT_SCH_MERGE_FETCH || msgType == TDMT_SCH_QUERY_HEARTBEAT || msgType == TDMT_SCH_DROP_TASK) {
return false;
}
return true;

7
source/client/src/clientImpl.c
View File

@ -590,6 +590,11 @@ int32_t buildAsyncExecNodeList(SRequestObj* pRequest, SArray** pNodeList, SArray
return code;
}
void freeVgList(void *list) {
SArray* pList = *(SArray**)list;
taosArrayDestroy(pList);
}
int32_t buildSyncExecNodeList(SRequestObj* pRequest, SArray** pNodeList, SArray* pMnodeList) {
SArray* pDbVgList = NULL;
SArray* pQnodeList = NULL;
@ -641,7 +646,7 @@ int32_t buildSyncExecNodeList(SRequestObj* pRequest, SArray** pNodeList, SArray*
_return:
taosArrayDestroy(pDbVgList);
taosArrayDestroyEx(pDbVgList, freeVgList);
taosArrayDestroy(pQnodeList);
return code;

29
source/client/src/clientStmt.c
View File

@ -6,11 +6,16 @@
#include "clientStmt.h"
static int32_t stmtCreateRequest(STscStmt* pStmt) {
int32_t code = 0;
if (pStmt->exec.pRequest == NULL) {
return buildRequest(pStmt->taos->id, pStmt->sql.sqlStr, pStmt->sql.sqlLen, NULL, false, &pStmt->exec.pRequest);
} else {
return TSDB_CODE_SUCCESS;
code = buildRequest(pStmt->taos->id, pStmt->sql.sqlStr, pStmt->sql.sqlLen, NULL, false, &pStmt->exec.pRequest);
if (TSDB_CODE_SUCCESS == code) {
pStmt->exec.pRequest->syncQuery = true;
}
}
return code;
}
int32_t stmtSwitchStatus(STscStmt* pStmt, STMT_STATUS newStatus) {
@ -155,7 +160,7 @@ int32_t stmtUpdateBindInfo(TAOS_STMT* stmt, STableMeta* pTableMeta, void* tags,
int32_t stmtUpdateExecInfo(TAOS_STMT* stmt, SHashObj* pVgHash, SHashObj* pBlockHash, bool autoCreateTbl) {
STscStmt* pStmt = (STscStmt*)stmt;
pStmt->exec.pVgHash = pVgHash;
pStmt->sql.pVgHash = pVgHash;
pStmt->exec.pBlockHash = pBlockHash;
pStmt->exec.autoCreateTbl = autoCreateTbl;
@ -177,7 +182,7 @@ int32_t stmtUpdateInfo(TAOS_STMT* stmt, STableMeta* pTableMeta, void* tags, char
int32_t stmtGetExecInfo(TAOS_STMT* stmt, SHashObj** pVgHash, SHashObj** pBlockHash) {
STscStmt* pStmt = (STscStmt*)stmt;
*pVgHash = pStmt->exec.pVgHash;
*pVgHash = pStmt->sql.pVgHash;
*pBlockHash = pStmt->exec.pBlockHash;
return TSDB_CODE_SUCCESS;
@ -227,7 +232,7 @@ int32_t stmtParseSql(STscStmt* pStmt) {
};
STMT_ERR_RET(stmtCreateRequest(pStmt));
STMT_ERR_RET(parseSql(pStmt->exec.pRequest, false, &pStmt->sql.pQuery, &stmtCb));
pStmt->bInfo.needParse = false;
@ -308,6 +313,8 @@ int32_t stmtCleanSQLInfo(STscStmt* pStmt) {
taosMemoryFree(pStmt->sql.sqlStr);
qDestroyQuery(pStmt->sql.pQuery);
taosArrayDestroy(pStmt->sql.nodeList);
taosHashCleanup(pStmt->sql.pVgHash);
pStmt->sql.pVgHash = NULL;
void* pIter = taosHashIterate(pStmt->sql.pTableCache, NULL);
while (pIter) {
@ -340,7 +347,7 @@ int32_t stmtRebuildDataBlock(STscStmt* pStmt, STableDataBlocks* pDataBlock, STab
STMT_ERR_RET(catalogGetTableHashVgroup(pStmt->pCatalog, &conn, &pStmt->bInfo.sname, &vgInfo));
STMT_ERR_RET(
taosHashPut(pStmt->exec.pVgHash, (const char*)&vgInfo.vgId, sizeof(vgInfo.vgId), (char*)&vgInfo, sizeof(vgInfo)));
taosHashPut(pStmt->sql.pVgHash, (const char*)&vgInfo.vgId, sizeof(vgInfo.vgId), (char*)&vgInfo, sizeof(vgInfo)));
STMT_ERR_RET(qRebuildStmtDataBlock(newBlock, pDataBlock, uid, vgInfo.vgId));
@ -680,6 +687,7 @@ int stmtBindBatch(TAOS_STMT* stmt, TAOS_MULTI_BIND* bind, int32_t colIdx) {
if (pStmt->sql.pQuery->haveResultSet) {
setResSchemaInfo(&pStmt->exec.pRequest->body.resInfo, pStmt->sql.pQuery->pResSchema,
pStmt->sql.pQuery->numOfResCols);
taosMemoryFreeClear(pStmt->sql.pQuery->pResSchema);
setResPrecision(&pStmt->exec.pRequest->body.resInfo, pStmt->sql.pQuery->precision);
}
@ -804,7 +812,7 @@ int stmtExec(TAOS_STMT* stmt) {
if (STMT_TYPE_QUERY == pStmt->sql.type) {
launchQueryImpl(pStmt->exec.pRequest, pStmt->sql.pQuery, true, NULL);
} else {
STMT_ERR_RET(qBuildStmtOutput(pStmt->sql.pQuery, pStmt->exec.pVgHash, pStmt->exec.pBlockHash));
STMT_ERR_RET(qBuildStmtOutput(pStmt->sql.pQuery, pStmt->sql.pVgHash, pStmt->exec.pBlockHash));
launchQueryImpl(pStmt->exec.pRequest, pStmt->sql.pQuery, true, (autoCreateTbl ? (void**)&pRsp : NULL));
}
@ -847,9 +855,10 @@ _return:
int stmtClose(TAOS_STMT* stmt) {
STscStmt* pStmt = (STscStmt*)stmt;
STMT_RET(stmtCleanSQLInfo(pStmt));
stmtCleanSQLInfo(pStmt);
taosMemoryFree(stmt);
return TSDB_CODE_SUCCESS;
}
const char* stmtErrstr(TAOS_STMT* stmt) {

2
source/client/test/clientTests.cpp
View File

@ -826,7 +826,7 @@ TEST(testCase, update_test) {
TAOS* pConn = taos_connect("localhost", "root", "taosdata", NULL, 0);
ASSERT_NE(pConn, nullptr);
TAOS_RES* pRes = taos_query(pConn, "create database if not exists abc1");
TAOS_RES* pRes = taos_query(pConn, "select cast(0 as timestamp)-1y");
if (taos_errno(pRes) != TSDB_CODE_SUCCESS) {
printf("failed to create database, code:%s", taos_errstr(pRes));
taos_free_result(pRes);

11
source/common/src/tdatablock.c
View File

@ -1107,6 +1107,11 @@ int32_t blockDataSort_rv(SSDataBlock* pDataBlock, SArray* pOrderInfo, bool nullF
void blockDataCleanup(SSDataBlock* pDataBlock) {
pDataBlock->info.rows = 0;
pDataBlock->info.groupId = 0;
pDataBlock->info.window.ekey = 0;
pDataBlock->info.window.skey = 0;
size_t numOfCols = taosArrayGetSize(pDataBlock->pDataBlock);
for (int32_t i = 0; i < numOfCols; ++i) {
SColumnInfoData* p = taosArrayGet(pDataBlock->pDataBlock, i);
@ -1163,9 +1168,15 @@ static int32_t doEnsureCapacity(SColumnInfoData* pColumn, const SDataBlockInfo*
void colInfoDataCleanup(SColumnInfoData* pColumn, uint32_t numOfRows) {
if (IS_VAR_DATA_TYPE(pColumn->info.type)) {
pColumn->varmeta.length = 0;
if (pColumn->varmeta.offset > 0) {
memset(pColumn->varmeta.offset, 0, sizeof(int32_t) * numOfRows);
}
} else {
if (pColumn->nullbitmap != NULL) {
memset(pColumn->nullbitmap, 0, BitmapLen(numOfRows));
if (pColumn->pData != NULL) {
memset(pColumn->pData, 0, pColumn->info.bytes * numOfRows);
}
}
}
}

18
source/common/src/tglobal.c
View File

@ -51,15 +51,15 @@ int32_t tsNumOfShmThreads = 1;
int32_t tsNumOfRpcThreads = 1;
int32_t tsNumOfCommitThreads = 2;
int32_t tsNumOfTaskQueueThreads = 1;
int32_t tsNumOfMnodeQueryThreads = 2;
int32_t tsNumOfMnodeQueryThreads = 4;
int32_t tsNumOfMnodeFetchThreads = 1;
int32_t tsNumOfMnodeReadThreads = 1;
int32_t tsNumOfVnodeQueryThreads = 2;
int32_t tsNumOfVnodeQueryThreads = 4;
int32_t tsNumOfVnodeStreamThreads = 2;
int32_t tsNumOfVnodeFetchThreads = 4;
int32_t tsNumOfVnodeWriteThreads = 2;
int32_t tsNumOfVnodeSyncThreads = 2;
int32_t tsNumOfQnodeQueryThreads = 2;
int32_t tsNumOfQnodeQueryThreads = 4;
int32_t tsNumOfQnodeFetchThreads = 4;
int32_t tsNumOfSnodeSharedThreads = 2;
int32_t tsNumOfSnodeUniqueThreads = 2;
@ -402,16 +402,16 @@ static int32_t taosAddServerCfg(SConfig *pCfg) {
tsNumOfCommitThreads = TRANGE(tsNumOfCommitThreads, 2, 4);
if (cfgAddInt32(pCfg, "numOfCommitThreads", tsNumOfCommitThreads, 1, 1024, 0) != 0) return -1;
tsNumOfMnodeQueryThreads = tsNumOfCores / 8;
tsNumOfMnodeQueryThreads = TRANGE(tsNumOfMnodeQueryThreads, 1, 4);
tsNumOfMnodeQueryThreads = tsNumOfCores * 2;
tsNumOfMnodeQueryThreads = TRANGE(tsNumOfMnodeQueryThreads, 4, 8);
if (cfgAddInt32(pCfg, "numOfMnodeQueryThreads", tsNumOfMnodeQueryThreads, 1, 1024, 0) != 0) return -1;
tsNumOfMnodeReadThreads = tsNumOfCores / 8;
tsNumOfMnodeReadThreads = TRANGE(tsNumOfMnodeReadThreads, 1, 4);
if (cfgAddInt32(pCfg, "numOfMnodeReadThreads", tsNumOfMnodeReadThreads, 1, 1024, 0) != 0) return -1;
tsNumOfVnodeQueryThreads = tsNumOfCores / 4;
tsNumOfVnodeQueryThreads = TMAX(tsNumOfVnodeQueryThreads, 2);
tsNumOfVnodeQueryThreads = tsNumOfCores * 2;
tsNumOfVnodeQueryThreads = TMAX(tsNumOfVnodeQueryThreads, 4);
if (cfgAddInt32(pCfg, "numOfVnodeQueryThreads", tsNumOfVnodeQueryThreads, 1, 1024, 0) != 0) return -1;
tsNumOfVnodeStreamThreads = tsNumOfCores / 4;
@ -430,8 +430,8 @@ static int32_t taosAddServerCfg(SConfig *pCfg) {
tsNumOfVnodeSyncThreads = TMAX(tsNumOfVnodeSyncThreads, 1);
if (cfgAddInt32(pCfg, "numOfVnodeSyncThreads", tsNumOfVnodeSyncThreads, 1, 1024, 0) != 0) return -1;
tsNumOfQnodeQueryThreads = tsNumOfCores / 2;
tsNumOfQnodeQueryThreads = TMAX(tsNumOfQnodeQueryThreads, 1);
tsNumOfQnodeQueryThreads = tsNumOfCores * 2;
tsNumOfQnodeQueryThreads = TMAX(tsNumOfQnodeQueryThreads, 4);
if (cfgAddInt32(pCfg, "numOfQnodeQueryThreads", tsNumOfQnodeQueryThreads, 1, 1024, 0) != 0) return -1;
tsNumOfQnodeFetchThreads = tsNumOfCores / 2;

6
source/common/src/tmsg.c
View File

@ -4941,14 +4941,14 @@ int tDecodeSVCreateStbReq(SDecoder *pCoder, SVCreateStbReq *pReq) {
return 0;
}
STSchema *tdGetSTSChemaFromSSChema(SSchema **pSchema, int32_t nCols) {
STSchema *tdGetSTSChemaFromSSChema(SSchema *pSchema, int32_t nCols, int32_t sver) {
STSchemaBuilder schemaBuilder = {0};
if (tdInitTSchemaBuilder(&schemaBuilder, 1) < 0) {
if (tdInitTSchemaBuilder(&schemaBuilder, sver) < 0) {
return NULL;
}
for (int i = 0; i < nCols; i++) {
SSchema *schema = *pSchema + i;
SSchema *schema = pSchema + i;
if (tdAddColToSchema(&schemaBuilder, schema->type, schema->flags, schema->colId, schema->bytes) < 0) {
tdDestroyTSchemaBuilder(&schemaBuilder);
return NULL;

5
source/common/src/tname.c
View File

@ -190,6 +190,11 @@ int32_t tNameGetDbName(const SName* name, char* dst) {
return 0;
}
const char* tNameGetDbNameP(const SName* name) {
return &name->dbname[0];
}
int32_t tNameGetFullDbName(const SName* name, char* dst) {
assert(name != NULL && dst != NULL);
snprintf(dst, TSDB_DB_FNAME_LEN, "%d.%s", name->acctId, name->dbname);

15
source/common/src/trow.c
View File

@ -568,6 +568,7 @@ int32_t tdSTSRowNew(SArray *pArray, STSchema *pTSchema, STSRow **ppRow) {
int32_t maxVarDataLen = 0;
int32_t iColVal = 0;
void *varBuf = NULL;
bool isAlloc = false;
ASSERT(nColVal > 1);
@ -610,8 +611,11 @@ int32_t tdSTSRowNew(SArray *pArray, STSchema *pTSchema, STSRow **ppRow) {
++iColVal;
}
*ppRow = (STSRow *)taosMemoryCalloc(
1, sizeof(STSRow) + pTSchema->flen + varDataLen + TD_BITMAP_BYTES(pTSchema->numOfCols - 1));
if (!(*ppRow)) {
*ppRow = (STSRow *)taosMemoryCalloc(
1, sizeof(STSRow) + pTSchema->flen + varDataLen + TD_BITMAP_BYTES(pTSchema->numOfCols - 1));
isAlloc = true;
}
if (!(*ppRow)) {
terrno = TSDB_CODE_OUT_OF_MEMORY;
@ -621,7 +625,9 @@ int32_t tdSTSRowNew(SArray *pArray, STSchema *pTSchema, STSRow **ppRow) {
if (maxVarDataLen > 0) {
varBuf = taosMemoryMalloc(maxVarDataLen);
if (!varBuf) {
taosMemoryFreeClear(*ppRow);
if(isAlloc) {
taosMemoryFreeClear(*ppRow);
}
terrno = TSDB_CODE_OUT_OF_MEMORY;
return -1;
}
@ -1323,12 +1329,11 @@ void tTSRowGetVal(STSRow *pRow, STSchema *pTSchema, int16_t iCol, SColVal *pColV
SCellVal cv;
SValue value;
ASSERT(iCol > 0);
ASSERT((pTColumn->colId == PRIMARYKEY_TIMESTAMP_COL_ID) || (iCol > 0));
if (TD_IS_TP_ROW(pRow)) {
tdSTpRowGetVal(pRow, pTColumn->colId, pTColumn->type, pTSchema->flen, pTColumn->offset, iCol - 1, &cv);
} else if (TD_IS_KV_ROW(pRow)) {
ASSERT(iCol > 0);
tdSKvRowGetVal(pRow, pTColumn->colId, iCol - 1, &cv);
} else {
ASSERT(0);

2
source/common/test/dataformatTest.cpp
View File

@ -116,7 +116,7 @@ STSchema *genSTSchema(int16_t nCols) {
}
STSchema *pResult = NULL;
pResult = tdGetSTSChemaFromSSChema(&pSchema, nCols);
pResult = tdGetSTSChemaFromSSChema(pSchema, nCols, 1);
taosMemoryFree(pSchema);
return pResult;

4
source/dnode/mgmt/exe/dmMain.c
View File

@ -158,8 +158,8 @@ static void taosCleanupArgs() {
}
int main(int argc, char const *argv[]) {
if (!taosCheckSystemIsSmallEnd()) {
printf("failed to start since on non-small-end machines\n");
if (!taosCheckSystemIsLittleEnd()) {
printf("failed to start since on non-little-end machines\n");
return -1;
}

5
source/dnode/mnode/impl/src/mndSubscribe.c
View File

@ -868,7 +868,10 @@ int32_t mndDropSubByTopic(SMnode *pMnode, STrans *pTrans, const char *topicName)
}
// iter all vnode to delete handle
ASSERT(taosHashGetSize(pSub->consumerHash) == 0);
if (taosHashGetSize(pSub->consumerHash) != 0) {
sdbRelease(pSdb, pSub);
return -1;
}
int32_t sz = taosArrayGetSize(pSub->unassignedVgs);
for (int32_t i = 0; i < sz; i++) {
SMqVgEp *pVgEp = taosArrayGetP(pSub->unassignedVgs, i);

9
source/dnode/mnode/impl/src/mndTopic.c
View File

@ -583,6 +583,7 @@ static int32_t mndProcessDropTopicReq(SRpcMsg *pReq) {
mndTransSetDbName(pTrans, pTopic->db, NULL);
if (pTrans == NULL) {
mError("topic:%s, failed to drop since %s", pTopic->name, terrstr());
mndReleaseTopic(pMnode, pTopic);
return -1;
}
@ -590,11 +591,17 @@ static int32_t mndProcessDropTopicReq(SRpcMsg *pReq) {
if (mndDropOffsetByTopic(pMnode, pTrans, dropReq.name) < 0) {
ASSERT(0);
mndTransDrop(pTrans);
mndReleaseTopic(pMnode, pTopic);
return -1;
}
// TODO check if rebalancing
if (mndDropSubByTopic(pMnode, pTrans, dropReq.name) < 0) {
ASSERT(0);
/*ASSERT(0);*/
mError("topic:%s, failed to drop since %s", pTopic->name, terrstr());
mndTransDrop(pTrans);
mndReleaseTopic(pMnode, pTopic);
return -1;
}

10
source/dnode/vnode/src/inc/tq.h
View File

@ -68,7 +68,7 @@ typedef struct {
typedef struct {
char* qmsg;
qTaskInfo_t task[5];
qTaskInfo_t task;
} STqExecCol;
typedef struct {
@ -82,13 +82,14 @@ typedef struct {
typedef struct {
int8_t subType;
STqReader* pExecReader[5];
STqReader* pExecReader;
union {
STqExecCol execCol;
STqExecTb execTb;
STqExecDb execDb;
};
int32_t numOfCols; // number of out pout column, temporarily used
int32_t numOfCols; // number of out pout column, temporarily used
SSchemaWrapper *pSchemaWrapper; // columns that are involved in query
} STqExecHandle;
typedef struct {
@ -138,8 +139,7 @@ int64_t tqScan(STQ* pTq, const STqHandle* pHandle, SMqDataRsp* pRsp, STqOffsetVa
int64_t tqFetchLog(STQ* pTq, STqHandle* pHandle, int64_t* fetchOffset, SWalCkHead** pHeadWithCkSum);
// tqExec
int32_t tqLogScanExec(STQ* pTq, STqExecHandle* pExec, SSubmitReq* pReq, SMqDataRsp* pRsp, int32_t workerId);
int32_t tqScanSnapshot(STQ* pTq, const STqExecHandle* pExec, SMqDataRsp* pRsp, STqOffsetVal offset, int32_t workerId);
int32_t tqLogScanExec(STQ* pTq, STqExecHandle* pExec, SSubmitReq* pReq, SMqDataRsp* pRsp);
int32_t tqSendDataRsp(STQ* pTq, const SRpcMsg* pMsg, const SMqPollReq* pReq, const SMqDataRsp* pRsp);
// tqMeta

135
source/dnode/vnode/src/inc/tsdb.h
View File

@ -40,12 +40,9 @@ typedef struct SDelIdx SDelIdx;
typedef struct STbData STbData;
typedef struct SMemTable SMemTable;
typedef struct STbDataIter STbDataIter;
typedef struct STable STable;
typedef struct SMapData SMapData;
typedef struct SBlockIdx SBlockIdx;
typedef struct SBlock SBlock;
typedef struct SBlockStatis SBlockStatis;
typedef struct SAggrBlkCol SAggrBlkCol;
typedef struct SColData SColData;
typedef struct SBlockDataHdr SBlockDataHdr;
typedef struct SBlockData SBlockData;
@ -62,8 +59,7 @@ typedef struct SDelFReader SDelFReader;
typedef struct SRowIter SRowIter;
typedef struct STsdbFS STsdbFS;
typedef struct SRowMerger SRowMerger;
typedef struct STsdbFSState STsdbFSState;
typedef struct STsdbSnapHdr STsdbSnapHdr;
typedef struct STsdbReadSnap STsdbReadSnap;
#define TSDB_MAX_SUBBLOCKS 8
#define TSDB_FHDR_SIZE 512
@ -176,8 +172,6 @@ void tsdbMemTableDestroy(SMemTable *pMemTable);
void tsdbGetTbDataFromMemTable(SMemTable *pMemTable, tb_uid_t suid, tb_uid_t uid, STbData **ppTbData);
void tsdbRefMemTable(SMemTable *pMemTable);
void tsdbUnrefMemTable(SMemTable *pMemTable);
int32_t tsdbTakeMemSnapshot(STsdb *pTsdb, SMemTable **ppMem, SMemTable **ppIMem);
void tsdbUntakeMemSnapshot(STsdb *pTsdb, SMemTable *pMem, SMemTable *pIMem);
// STbDataIter
int32_t tsdbTbDataIterCreate(STbData *pTbData, TSDBKEY *pFrom, int8_t backward, STbDataIter **ppIter);
void *tsdbTbDataIterDestroy(STbDataIter *pIter);
@ -188,30 +182,39 @@ bool tsdbTbDataIterNext(STbDataIter *pIter);
int32_t tsdbGetNRowsInTbData(STbData *pTbData);
// tsdbFile.c ==============================================================================================
typedef enum { TSDB_HEAD_FILE = 0, TSDB_DATA_FILE, TSDB_LAST_FILE, TSDB_SMA_FILE } EDataFileT;
void tsdbDataFileName(STsdb *pTsdb, SDFileSet *pDFileSet, EDataFileT ftype, char fname[]);
bool tsdbFileIsSame(SDFileSet *pDFileSet1, SDFileSet *pDFileSet2, EDataFileT ftype);
bool tsdbDelFileIsSame(SDelFile *pDelFile1, SDelFile *pDelFile2);
int32_t tsdbUpdateDFileHdr(TdFilePtr pFD, SDFileSet *pSet, EDataFileT ftype);
int32_t tsdbDFileRollback(STsdb *pTsdb, SDFileSet *pSet, EDataFileT ftype);
int32_t tPutDataFileHdr(uint8_t *p, SDFileSet *pSet, EDataFileT ftype);
int32_t tPutHeadFile(uint8_t *p, SHeadFile *pHeadFile);
int32_t tPutDataFile(uint8_t *p, SDataFile *pDataFile);
int32_t tPutLastFile(uint8_t *p, SLastFile *pLastFile);
int32_t tPutSmaFile(uint8_t *p, SSmaFile *pSmaFile);
int32_t tPutDelFile(uint8_t *p, SDelFile *pDelFile);
int32_t tGetDelFile(uint8_t *p, SDelFile *pDelFile);
int32_t tPutDFileSet(uint8_t *p, SDFileSet *pSet);
int32_t tGetDFileSet(uint8_t *p, SDFileSet *pSet);
void tsdbHeadFileName(STsdb *pTsdb, SDiskID did, int32_t fid, SHeadFile *pHeadF, char fname[]);
void tsdbDataFileName(STsdb *pTsdb, SDiskID did, int32_t fid, SDataFile *pDataF, char fname[]);
void tsdbLastFileName(STsdb *pTsdb, SDiskID did, int32_t fid, SLastFile *pLastF, char fname[]);
void tsdbSmaFileName(STsdb *pTsdb, SDiskID did, int32_t fid, SSmaFile *pSmaF, char fname[]);
// SDelFile
void tsdbDelFileName(STsdb *pTsdb, SDelFile *pFile, char fname[]);
// tsdbFS.c ==============================================================================================
int32_t tsdbFSOpen(STsdb *pTsdb, STsdbFS **ppFS);
int32_t tsdbFSClose(STsdbFS *pFS);
int32_t tsdbFSBegin(STsdbFS *pFS);
int32_t tsdbFSCommit(STsdbFS *pFS);
int32_t tsdbFSOpen(STsdb *pTsdb);
int32_t tsdbFSClose(STsdb *pTsdb);
int32_t tsdbFSCopy(STsdb *pTsdb, STsdbFS *pFS);
void tsdbFSDestroy(STsdbFS *pFS);
int32_t tDFileSetCmprFn(const void *p1, const void *p2);
int32_t tsdbFSCommit1(STsdb *pTsdb, STsdbFS *pFS);
int32_t tsdbFSCommit2(STsdb *pTsdb, STsdbFS *pFS);
int32_t tsdbFSRef(STsdb *pTsdb, STsdbFS *pFS);
void tsdbFSUnref(STsdb *pTsdb, STsdbFS *pFS);
int32_t tsdbFSRollback(STsdbFS *pFS);
int32_t tsdbFSStateUpsertDelFile(STsdbFSState *pState, SDelFile *pDelFile);
int32_t tsdbFSStateUpsertDFileSet(STsdbFSState *pState, SDFileSet *pSet);
void tsdbFSStateDeleteDFileSet(STsdbFSState *pState, int32_t fid);
SDelFile *tsdbFSStateGetDelFile(STsdbFSState *pState);
SDFileSet *tsdbFSStateGetDFileSet(STsdbFSState *pState, int32_t fid, int32_t flag);
int32_t tsdbFSUpsertFSet(STsdbFS *pFS, SDFileSet *pSet);
int32_t tsdbFSUpsertDelFile(STsdbFS *pFS, SDelFile *pDelFile);
// tsdbReaderWriter.c ==============================================================================================
// SDataFWriter
int32_t tsdbDataFWriterOpen(SDataFWriter **ppWriter, STsdb *pTsdb, SDFileSet *pSet);
@ -222,8 +225,7 @@ int32_t tsdbWriteBlock(SDataFWriter *pWriter, SMapData *pMapData, uint8_t **ppBu
int32_t tsdbWriteBlockData(SDataFWriter *pWriter, SBlockData *pBlockData, uint8_t **ppBuf1, uint8_t **ppBuf2,
SBlockIdx *pBlockIdx, SBlock *pBlock, int8_t cmprAlg);
SDFileSet *tsdbDataFWriterGetWSet(SDataFWriter *pWriter);
int32_t tsdbDFileSetCopy(STsdb *pTsdb, SDFileSet *pSetFrom, SDFileSet *pSetTo);
int32_t tsdbDFileSetCopy(STsdb *pTsdb, SDFileSet *pSetFrom, SDFileSet *pSetTo);
// SDataFReader
int32_t tsdbDataFReaderOpen(SDataFReader **ppReader, STsdb *pTsdb, SDFileSet *pSet);
int32_t tsdbDataFReaderClose(SDataFReader **ppReader);
@ -245,6 +247,9 @@ int32_t tsdbDelFReaderOpen(SDelFReader **ppReader, SDelFile *pFile, STsdb *pTsdb
int32_t tsdbDelFReaderClose(SDelFReader **ppReader);
int32_t tsdbReadDelData(SDelFReader *pReader, SDelIdx *pDelIdx, SArray *aDelData, uint8_t **ppBuf);
int32_t tsdbReadDelIdx(SDelFReader *pReader, SArray *aDelIdx, uint8_t **ppBuf);
// tsdbRead.c ==============================================================================================
int32_t tsdbTakeReadSnap(STsdb *pTsdb, STsdbReadSnap **ppSnap);
void tsdbUntakeReadSnap(STsdb *pTsdb, STsdbReadSnap *pSnap);
#define TSDB_CACHE_NO(c) ((c).cacheLast == 0)
#define TSDB_CACHE_LAST_ROW(c) (((c).cacheLast & 1) > 0)
@ -276,6 +281,11 @@ typedef struct {
TSKEY minKey;
} SRtn;
struct STsdbFS {
SDelFile *pDelFile;
SArray *aDFileSet; // SArray<SDFileSet>
};
struct STsdb {
char *path;
SVnode *pVnode;
@ -283,7 +293,7 @@ struct STsdb {
TdThreadRwlock rwLock;
SMemTable *mem;
SMemTable *imem;
STsdbFS *pFS;
STsdbFS fs;
SLRUCache *lruCache;
};
@ -402,16 +412,6 @@ struct SBlock {
SSubBlock aSubBlock[TSDB_MAX_SUBBLOCKS];
};
struct SAggrBlkCol {
int16_t colId;
int16_t maxIndex;
int16_t minIndex;
int16_t numOfNull;
int64_t sum;
int64_t max;
int64_t min;
};
struct SColData {
int16_t cid;
int8_t type;
@ -465,12 +465,6 @@ struct SDelIdx {
int64_t size;
};
struct SDelFile {
int64_t commitID;
int64_t size;
int64_t offset;
};
#pragma pack(push, 1)
struct SBlockDataHdr {
uint32_t delimiter;
@ -479,34 +473,50 @@ struct SBlockDataHdr {
};
#pragma pack(pop)
struct SDelFile {
volatile int32_t nRef;
int64_t commitID;
int64_t size;
int64_t offset;
};
struct SHeadFile {
volatile int32_t nRef;
int64_t commitID;
int64_t size;
int64_t offset;
};
struct SDataFile {
volatile int32_t nRef;
int64_t commitID;
int64_t size;
};
struct SLastFile {
volatile int32_t nRef;
int64_t commitID;
int64_t size;
};
struct SSmaFile {
volatile int32_t nRef;
int64_t commitID;
int64_t size;
};
struct SDFileSet {
SDiskID diskId;
int32_t fid;
SHeadFile fHead;
SDataFile fData;
SLastFile fLast;
SSmaFile fSma;
SDiskID diskId;
int32_t fid;
SHeadFile *pHeadF;
SDataFile *pDataF;
SLastFile *pLastF;
SSmaFile *pSmaF;
};
struct SRowIter {
@ -521,26 +531,33 @@ struct SRowMerger {
SArray *pArray; // SArray<SColVal>
};
struct STsdbFSState {
SDelFile *pDelFile;
SArray *aDFileSet; // SArray<aDFileSet>
SDelFile delFile;
};
struct STsdbFS {
STsdb *pTsdb;
TdThreadRwlock lock;
int8_t inTxn;
STsdbFSState *cState;
STsdbFSState *nState;
};
struct SDelFWriter {
STsdb *pTsdb;
SDelFile fDel;
TdFilePtr pWriteH;
};
struct SDataFWriter {
STsdb *pTsdb;
SDFileSet wSet;
TdFilePtr pHeadFD;
TdFilePtr pDataFD;
TdFilePtr pLastFD;
TdFilePtr pSmaFD;
SHeadFile fHead;
SDataFile fData;
SLastFile fLast;
SSmaFile fSma;
};
struct STsdbReadSnap {
SMemTable *pMem;
SMemTable *pIMem;
STsdbFS fs;
};
#ifdef __cplusplus
}
#endif

2
source/dnode/vnode/src/inc/vnodeInt.h
View File

@ -146,7 +146,7 @@ int32_t tqCheckColModifiable(STQ* pTq, int32_t colId);
int32_t tqProcessVgChangeReq(STQ* pTq, char* msg, int32_t msgLen);
int32_t tqProcessVgDeleteReq(STQ* pTq, char* msg, int32_t msgLen);
int32_t tqProcessOffsetCommitReq(STQ* pTq, char* msg, int32_t msgLen);
int32_t tqProcessPollReq(STQ* pTq, SRpcMsg* pMsg, int32_t workerId);
int32_t tqProcessPollReq(STQ* pTq, SRpcMsg* pMsg);
int32_t tqProcessTaskDeployReq(STQ* pTq, char* msg, int32_t msgLen);
int32_t tqProcessTaskDropReq(STQ* pTq, char* msg, int32_t msgLen);
int32_t tqProcessStreamTrigger(STQ* pTq, SSubmitReq* data);

57
source/dnode/vnode/src/sma/smaRollup.c
View File

@ -579,9 +579,11 @@ static int32_t tdRSmaFetchAndSubmitResult(SRSmaInfoItem *pItem, STSchema *pTSche
while (1) {
SSDataBlock *output = NULL;
uint64_t ts;
if (qExecTask(pItem->taskInfo, &output, &ts) < 0) {
smaError("vgId:%d, qExecTask for rsma table %" PRIi64 "l evel %" PRIi8 " failed since %s", SMA_VID(pSma), suid,
pItem->level, terrstr());
int32_t code = qExecTask(pItem->taskInfo, &output, &ts);
if (code < 0) {
smaError("vgId:%d, qExecTask for rsma table %" PRIi64 " level %" PRIi8 " failed since %s", SMA_VID(pSma), suid,
pItem->level, terrstr(code));
goto _err;
}
if (!output) {
@ -597,35 +599,36 @@ static int32_t tdRSmaFetchAndSubmitResult(SRSmaInfoItem *pItem, STSchema *pTSche
}
taosArrayPush(pResult, output);
}
if (taosArrayGetSize(pResult) > 0) {
if (taosArrayGetSize(pResult) > 0) {
#if 1
char flag[10] = {0};
snprintf(flag, 10, "level %" PRIi8, pItem->level);
blockDebugShowDataBlocks(pResult, flag);
char flag[10] = {0};
snprintf(flag, 10, "level %" PRIi8, pItem->level);
blockDebugShowDataBlocks(pResult, flag);
#endif
STsdb *sinkTsdb = (pItem->level == TSDB_RETENTION_L1 ? pSma->pRSmaTsdb[0] : pSma->pRSmaTsdb[1]);
SSubmitReq *pReq = NULL;
// TODO: the schema update should be handled
if (buildSubmitReqFromDataBlock(&pReq, pResult, pTSchema, SMA_VID(pSma), suid) < 0) {
smaError("vgId:%d, build submit req for rsma table %" PRIi64 "l evel %" PRIi8 " failed since %s", SMA_VID(pSma),
suid, pItem->level, terrstr());
goto _err;
}
STsdb *sinkTsdb = (pItem->level == TSDB_RETENTION_L1 ? pSma->pRSmaTsdb[0] : pSma->pRSmaTsdb[1]);
SSubmitReq *pReq = NULL;
// TODO: the schema update should be handled
if (buildSubmitReqFromDataBlock(&pReq, pResult, pTSchema, SMA_VID(pSma), suid) < 0) {
smaError("vgId:%d, build submit req for rsma table %" PRIi64 "l evel %" PRIi8 " failed since %s", SMA_VID(pSma),
suid, pItem->level, terrstr());
goto _err;
}
if (pReq && tdProcessSubmitReq(sinkTsdb, atomic_add_fetch_64(&pStat->submitVer, 1), pReq) < 0) {
taosMemoryFreeClear(pReq);
smaError("vgId:%d, process submit req for rsma table %" PRIi64 " level %" PRIi8 " failed since %s",
SMA_VID(pSma), suid, pItem->level, terrstr());
goto _err;
}
if (pReq && tdProcessSubmitReq(sinkTsdb, atomic_add_fetch_64(&pStat->submitVer, 1), pReq) < 0) {
taosMemoryFreeClear(pReq);
smaError("vgId:%d, process submit req for rsma table %" PRIi64 " level %" PRIi8 " failed since %s", SMA_VID(pSma),
suid, pItem->level, terrstr());
goto _err;
taosArrayClear(pResult);
} else if (terrno == 0) {
smaDebug("vgId:%d, no rsma %" PRIi8 " data fetched yet", SMA_VID(pSma), pItem->level);
} else {
smaDebug("vgId:%d, no rsma %" PRIi8 " data fetched since %s", SMA_VID(pSma), pItem->level, tstrerror(terrno));
}
taosMemoryFreeClear(pReq);
} else if (terrno == 0) {
smaDebug("vgId:%d, no rsma %" PRIi8 " data fetched yet", SMA_VID(pSma), pItem->level);
} else {
smaDebug("vgId:%d, no rsma %" PRIi8 " data fetched since %s", SMA_VID(pSma), pItem->level, tstrerror(terrno));
}
tdDestroySDataBlockArray(pResult);
@ -1364,4 +1367,4 @@ static void tdRSmaFetchTrigger(void *param, void *tmrId) {
_end:
tdReleaseSmaRef(smaMgmt.rsetId, pItem->refId, __func__, __LINE__);
}
}

58
source/dnode/vnode/src/tq/tq.c
View File

@ -215,10 +215,10 @@ int32_t tqCheckColModifiable(STQ* pTq, int32_t colId) {
if (pIter == NULL) break;
STqHandle* pExec = (STqHandle*)pIter;
if (pExec->execHandle.subType == TOPIC_SUB_TYPE__COLUMN) {
int32_t sz = taosArrayGetSize(pExec->colIdList);
int32_t sz = pExec->execHandle.pSchemaWrapper->nCols;
for (int32_t i = 0; i < sz; i++) {
int32_t forbidColId = *(int32_t*)taosArrayGet(pExec->colIdList, i);
if (forbidColId == colId) {
SSchema* pSchema = &pExec->execHandle.pSchemaWrapper->pSchema[i];
if (pSchema->colId == colId) {
taosHashCancelIterate(pTq->handles, pIter);
return -1;
}
@ -262,7 +262,7 @@ static int32_t tqInitDataRsp(SMqDataRsp* pRsp, const SMqPollReq* pReq, int8_t su
static int32_t tqInitMetaRsp(SMqMetaRsp* pRsp, const SMqPollReq* pReq) { return 0; }
int32_t tqProcessPollReq(STQ* pTq, SRpcMsg* pMsg, int32_t workerId) {
int32_t tqProcessPollReq(STQ* pTq, SRpcMsg* pMsg) {
SMqPollReq* pReq = pMsg->pCont;
int64_t consumerId = pReq->consumerId;
int64_t timeout = pReq->timeout;
@ -271,9 +271,6 @@ int32_t tqProcessPollReq(STQ* pTq, SRpcMsg* pMsg, int32_t workerId) {
STqOffsetVal reqOffset = pReq->reqOffset;
STqOffsetVal fetchOffsetNew;
// todo
workerId = 0;
// 1.find handle
STqHandle* pHandle = taosHashGet(pTq->handles, pReq->subKey, strlen(pReq->subKey));
/*ASSERT(pHandle);*/
@ -405,7 +402,7 @@ int32_t tqProcessPollReq(STQ* pTq, SRpcMsg* pMsg, int32_t workerId) {
if (pHead->msgType == TDMT_VND_SUBMIT) {
SSubmitReq* pCont = (SSubmitReq*)&pHead->body;
if (tqLogScanExec(pTq, &pHandle->execHandle, pCont, &dataRsp, workerId) < 0) {
if (tqLogScanExec(pTq, &pHandle->execHandle, pCont, &dataRsp) < 0) {
/*ASSERT(0);*/
}
// TODO batch optimization:
@ -518,27 +515,24 @@ int32_t tqProcessVgChangeReq(STQ* pTq, char* msg, int32_t msgLen) {
pHandle->execHandle.execCol.qmsg = req.qmsg;
pHandle->snapshotVer = ver;
req.qmsg = NULL;
for (int32_t i = 0; i < 5; i++) {
SReadHandle handle = {
.meta = pTq->pVnode->pMeta,
.vnode = pTq->pVnode,
.initTableReader = true,
.initTqReader = true,
.version = ver,
};
pHandle->execHandle.execCol.task[i] =
qCreateQueueExecTaskInfo(pHandle->execHandle.execCol.qmsg, &handle, &pHandle->execHandle.numOfCols);
ASSERT(pHandle->execHandle.execCol.task[i]);
void* scanner = NULL;
qExtractStreamScanner(pHandle->execHandle.execCol.task[i], &scanner);
ASSERT(scanner);
pHandle->execHandle.pExecReader[i] = qExtractReaderFromStreamScanner(scanner);
ASSERT(pHandle->execHandle.pExecReader[i]);
}
SReadHandle handle = {
.meta = pTq->pVnode->pMeta,
.vnode = pTq->pVnode,
.initTableReader = true,
.initTqReader = true,
.version = ver,
};
pHandle->execHandle.execCol.task =
qCreateQueueExecTaskInfo(pHandle->execHandle.execCol.qmsg, &handle, &pHandle->execHandle.numOfCols,
&pHandle->execHandle.pSchemaWrapper);
ASSERT(pHandle->execHandle.execCol.task);
void* scanner = NULL;
qExtractStreamScanner(pHandle->execHandle.execCol.task, &scanner);
ASSERT(scanner);
pHandle->execHandle.pExecReader = qExtractReaderFromStreamScanner(scanner);
ASSERT(pHandle->execHandle.pExecReader);
} else if (pHandle->execHandle.subType == TOPIC_SUB_TYPE__DB) {
for (int32_t i = 0; i < 5; i++) {
pHandle->execHandle.pExecReader[i] = tqOpenReader(pTq->pVnode);
}
pHandle->execHandle.pExecReader = tqOpenReader(pTq->pVnode);
pHandle->execHandle.execDb.pFilterOutTbUid =
taosHashInit(64, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BIGINT), false, HASH_NO_LOCK);
} else if (pHandle->execHandle.subType == TOPIC_SUB_TYPE__TABLE) {
@ -550,10 +544,8 @@ int32_t tqProcessVgChangeReq(STQ* pTq, char* msg, int32_t msgLen) {
int64_t tbUid = *(int64_t*)taosArrayGet(tbUidList, i);
tqDebug("vgId:%d, idx %d, uid:%" PRId64, TD_VID(pTq->pVnode), i, tbUid);
}
for (int32_t i = 0; i < 5; i++) {
pHandle->execHandle.pExecReader[i] = tqOpenReader(pTq->pVnode);
tqReaderSetTbUidList(pHandle->execHandle.pExecReader[i], tbUidList);
}
pHandle->execHandle.pExecReader = tqOpenReader(pTq->pVnode);
tqReaderSetTbUidList(pHandle->execHandle.pExecReader, tbUidList);
taosArrayDestroy(tbUidList);
}
taosHashPut(pTq->handles, req.subKey, strlen(req.subKey), pHandle, sizeof(STqHandle));
@ -634,7 +626,7 @@ int32_t tqProcessTaskDeployReq(STQ* pTq, char* msg, int32_t msgLen) {
ASSERT(pTask->tbSink.pSchemaWrapper->pSchema);
pTask->tbSink.pTSchema =
tdGetSTSChemaFromSSChema(&pTask->tbSink.pSchemaWrapper->pSchema, pTask->tbSink.pSchemaWrapper->nCols);
tdGetSTSChemaFromSSChema(pTask->tbSink.pSchemaWrapper->pSchema, pTask->tbSink.pSchemaWrapper->nCols, 1);
ASSERT(pTask->tbSink.pTSchema);
}

24
source/dnode/vnode/src/tq/tqExec.c
View File

@ -37,8 +37,8 @@ static int32_t tqAddBlockDataToRsp(const SSDataBlock* pBlock, SMqDataRsp* pRsp,
return 0;
}
static int32_t tqAddBlockSchemaToRsp(const STqExecHandle* pExec, int32_t workerId, SMqDataRsp* pRsp) {
SSchemaWrapper* pSW = tCloneSSchemaWrapper(pExec->pExecReader[workerId]->pSchemaWrapper);
static int32_t tqAddBlockSchemaToRsp(const STqExecHandle* pExec, SMqDataRsp* pRsp) {
SSchemaWrapper* pSW = tCloneSSchemaWrapper(pExec->pExecReader->pSchemaWrapper);
if (pSW == NULL) {
return -1;
}
@ -51,6 +51,7 @@ static int32_t tqAddTbNameToRsp(const STQ* pTq, int64_t uid, SMqDataRsp* pRsp) {
metaReaderInit(&mr, pTq->pVnode->pMeta, 0);
// TODO add reference to gurantee success
if (metaGetTableEntryByUid(&mr, uid) < 0) {
metaReaderClear(&mr);
return -1;
}
char* tbName = strdup(mr.me.name);
@ -61,7 +62,7 @@ static int32_t tqAddTbNameToRsp(const STQ* pTq, int64_t uid, SMqDataRsp* pRsp) {
int64_t tqScan(STQ* pTq, const STqHandle* pHandle, SMqDataRsp* pRsp, STqOffsetVal* pOffset) {
const STqExecHandle* pExec = &pHandle->execHandle;
qTaskInfo_t task = pExec->execCol.task[0];
qTaskInfo_t task = pExec->execCol.task;
if (qStreamPrepareScan(task, pOffset) < 0) {
if (pOffset->type == TMQ_OFFSET__LOG) {
@ -89,7 +90,7 @@ int64_t tqScan(STQ* pTq, const STqHandle* pHandle, SMqDataRsp* pRsp, STqOffsetVa
if (pDataBlock != NULL) {
if (pRsp->withTbName) {
if (pOffset->type == TMQ_OFFSET__LOG) {
int64_t uid = pExec->pExecReader[0]->msgIter.uid;
int64_t uid = pExec->pExecReader->msgIter.uid;
if (tqAddTbNameToRsp(pTq, uid, pRsp) < 0) {
continue;
}
@ -108,6 +109,7 @@ int64_t tqScan(STQ* pTq, const STqHandle* pHandle, SMqDataRsp* pRsp, STqOffsetVa
}
if (pRsp->blockNum == 0 && pOffset->type == TMQ_OFFSET__SNAPSHOT_DATA) {
tqDebug("vgId: %d, tsdb consume over, switch to wal, ver %ld", TD_VID(pTq->pVnode), pHandle->snapshotVer + 1);
tqOffsetResetToLog(pOffset, pHandle->snapshotVer);
qStreamPrepareScan(task, pOffset);
continue;
@ -184,12 +186,12 @@ int32_t tqScanSnapshot(STQ* pTq, const STqExecHandle* pExec, SMqDataRsp* pRsp, S
}
#endif
int32_t tqLogScanExec(STQ* pTq, STqExecHandle* pExec, SSubmitReq* pReq, SMqDataRsp* pRsp, int32_t workerId) {
int32_t tqLogScanExec(STQ* pTq, STqExecHandle* pExec, SSubmitReq* pReq, SMqDataRsp* pRsp) {
ASSERT(pExec->subType != TOPIC_SUB_TYPE__COLUMN);
if (pExec->subType == TOPIC_SUB_TYPE__TABLE) {
pRsp->withSchema = 1;
STqReader* pReader = pExec->pExecReader[workerId];
STqReader* pReader = pExec->pExecReader;
tqReaderSetDataMsg(pReader, pReq, 0);
while (tqNextDataBlock(pReader)) {
SSDataBlock block = {0};
@ -197,18 +199,18 @@ int32_t tqLogScanExec(STQ* pTq, STqExecHandle* pExec, SSubmitReq* pReq, SMqDataR
if (terrno == TSDB_CODE_TQ_TABLE_SCHEMA_NOT_FOUND) continue;
}
if (pRsp->withTbName) {
int64_t uid = pExec->pExecReader[workerId]->msgIter.uid;
int64_t uid = pExec->pExecReader->msgIter.uid;
if (tqAddTbNameToRsp(pTq, uid, pRsp) < 0) {
continue;
}
}
tqAddBlockDataToRsp(&block, pRsp, taosArrayGetSize(block.pDataBlock));
tqAddBlockSchemaToRsp(pExec, workerId, pRsp);
tqAddBlockSchemaToRsp(pExec, pRsp);
pRsp->blockNum++;
}
} else if (pExec->subType == TOPIC_SUB_TYPE__DB) {
pRsp->withSchema = 1;
STqReader* pReader = pExec->pExecReader[workerId];
STqReader* pReader = pExec->pExecReader;
tqReaderSetDataMsg(pReader, pReq, 0);
while (tqNextDataBlockFilterOut(pReader, pExec->execDb.pFilterOutTbUid)) {
SSDataBlock block = {0};
@ -216,13 +218,13 @@ int32_t tqLogScanExec(STQ* pTq, STqExecHandle* pExec, SSubmitReq* pReq, SMqDataR
if (terrno == TSDB_CODE_TQ_TABLE_SCHEMA_NOT_FOUND) continue;
}
if (pRsp->withTbName) {
int64_t uid = pExec->pExecReader[workerId]->msgIter.uid;
int64_t uid = pExec->pExecReader->msgIter.uid;
if (tqAddTbNameToRsp(pTq, uid, pRsp) < 0) {
continue;
}
}
tqAddBlockDataToRsp(&block, pRsp, taosArrayGetSize(block.pDataBlock));
tqAddBlockSchemaToRsp(pExec, workerId, pRsp);
tqAddBlockSchemaToRsp(pExec, pRsp);
pRsp->blockNum++;
}
}

34
source/dnode/vnode/src/tq/tqMeta.c
View File

@ -80,27 +80,23 @@ int32_t tqMetaOpen(STQ* pTq) {
tDecoderInit(&decoder, (uint8_t*)pVal, vLen);
tDecodeSTqHandle(&decoder, &handle);
handle.pWalReader = walOpenReader(pTq->pVnode->pWal, NULL);
/*for (int32_t i = 0; i < 5; i++) {*/
/*handle.execHandle.pExecReader[i] = tqOpenReader(pTq->pVnode);*/
/*}*/
if (handle.execHandle.subType == TOPIC_SUB_TYPE__COLUMN) {
for (int32_t i = 0; i < 5; i++) {
SReadHandle reader = {
.meta = pTq->pVnode->pMeta,
.vnode = pTq->pVnode,
.initTableReader = true,
.initTqReader = true,
.version = handle.snapshotVer,
};
SReadHandle reader = {
.meta = pTq->pVnode->pMeta,
.vnode = pTq->pVnode,
.initTableReader = true,
.initTqReader = true,
.version = handle.snapshotVer,
};
handle.execHandle.execCol.task[i] = qCreateQueueExecTaskInfo(handle.execHandle.execCol.qmsg, &reader, &handle.execHandle.numOfCols);
ASSERT(handle.execHandle.execCol.task[i]);
void* scanner = NULL;
qExtractStreamScanner(handle.execHandle.execCol.task[i], &scanner);
ASSERT(scanner);
handle.execHandle.pExecReader[i] = qExtractReaderFromStreamScanner(scanner);
ASSERT(handle.execHandle.pExecReader[i]);
}
handle.execHandle.execCol.task =
qCreateQueueExecTaskInfo(handle.execHandle.execCol.qmsg, &reader, &handle.execHandle.numOfCols, &handle.execHandle.pSchemaWrapper);
ASSERT(handle.execHandle.execCol.task);
void* scanner = NULL;
qExtractStreamScanner(handle.execHandle.execCol.task, &scanner);
ASSERT(scanner);
handle.execHandle.pExecReader = qExtractReaderFromStreamScanner(scanner);
ASSERT(handle.execHandle.pExecReader);
} else {
handle.execHandle.execDb.pFilterOutTbUid =
taosHashInit(64, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BIGINT), false, HASH_NO_LOCK);

2
source/dnode/vnode/src/tq/tqPush.c
View File

@ -249,6 +249,8 @@ int tqPushMsg(STQ* pTq, void* msg, int32_t msgLen, tmsg_t msgType, int64_t ver)
return -1;
}
memcpy(data, msg, msgLen);
SSubmitReq* pReq = (SSubmitReq*)data;
pReq->version = ver;
tqProcessStreamTrigger(pTq, data);
}

7
source/dnode/vnode/src/tq/tqRead.c
View File

@ -314,6 +314,7 @@ int32_t tqRetrieveDataBlock(SSDataBlock* pBlock, STqReader* pReader) {
pBlock->info.uid = pReader->msgIter.uid;
pBlock->info.rows = pReader->msgIter.numOfRows;
pBlock->info.version = pReader->pMsg->version;
while ((row = tGetSubmitBlkNext(&pReader->blkIter)) != NULL) {
tdSTSRowIterReset(&iter, row);
@ -393,10 +394,8 @@ int32_t tqUpdateTbUidList(STQ* pTq, const SArray* tbUidList, bool isAdd) {
if (pIter == NULL) break;
STqHandle* pExec = (STqHandle*)pIter;
if (pExec->execHandle.subType == TOPIC_SUB_TYPE__COLUMN) {
for (int32_t i = 0; i < 5; i++) {
int32_t code = qUpdateQualifiedTableId(pExec->execHandle.execCol.task[i], tbUidList, isAdd);
ASSERT(code == 0);
}
int32_t code = qUpdateQualifiedTableId(pExec->execHandle.execCol.task, tbUidList, isAdd);
ASSERT(code == 0);
} else if (pExec->execHandle.subType == TOPIC_SUB_TYPE__DB) {
if (!isAdd) {
int32_t sz = taosArrayGetSize(tbUidList);

7
source/dnode/vnode/src/tq/tqSink.c
View File

@ -127,6 +127,8 @@ SSubmitReq* tdBlockToSubmit(const SArray* pBlocks, const STSchema* pTSchema, boo
int32_t rows = pDataBlock->info.rows;
tqDebug("tq sink, convert block %d, rows: %d", i, rows);
int32_t dataLen = 0;
void* blkSchema = POINTER_SHIFT(blkHead, sizeof(SSubmitBlk));
@ -178,11 +180,14 @@ void tqTableSink(SStreamTask* pTask, void* vnode, int64_t ver, void* data) {
const SArray* pRes = (const SArray*)data;
SVnode* pVnode = (SVnode*)vnode;
tqDebug("task write into table, vgId %d, block num: %d", pVnode->config.vgId, (int32_t)pRes->size);
tqDebug("vgId:%d, task %d write into table, block num: %d", TD_VID(pVnode), pTask->taskId, (int32_t)pRes->size);
ASSERT(pTask->tbSink.pTSchema);
SSubmitReq* pReq = tdBlockToSubmit(pRes, pTask->tbSink.pTSchema, true, pTask->tbSink.stbUid,
pTask->tbSink.stbFullName, pVnode->config.vgId);
tqDebug("vgId:%d, task %d convert blocks over, put into write-queue", TD_VID(pVnode), pTask->taskId);
/*tPrintFixedSchemaSubmitReq(pReq, pTask->tbSink.pTSchema);*/
// build write msg
SRpcMsg msg = {

804
source/dnode/vnode/src/tsdb/tsdbCache.c
View File

@ -464,7 +464,7 @@ static int32_t getNextRowFromFS(void *iter, TSDBROW **ppRow) {
switch (state->state) {
case SFSNEXTROW_FS:
state->aDFileSet = state->pTsdb->pFS->cState->aDFileSet;
// state->aDFileSet = state->pTsdb->pFS->cState->aDFileSet;
state->nFileSet = taosArrayGetSize(state->aDFileSet);
state->iFileSet = state->nFileSet;
@ -793,9 +793,10 @@ typedef struct {
TSDBROW memRow, imemRow, fsRow;
TsdbNextRowState input[3];
SMemTable *pMemTable;
SMemTable *pIMemTable;
STsdb *pTsdb;
// SMemTable *pMemTable;
// SMemTable *pIMemTable;
STsdbReadSnap *pReadSnap;
STsdb *pTsdb;
} CacheNextRowIter;
static int32_t nextRowIterOpen(CacheNextRowIter *pIter, tb_uid_t uid, STsdb *pTsdb) {
@ -803,16 +804,16 @@ static int32_t nextRowIterOpen(CacheNextRowIter *pIter, tb_uid_t uid, STsdb *pTs
tb_uid_t suid = getTableSuidByUid(uid, pTsdb);
tsdbTakeMemSnapshot(pTsdb, &pIter->pMemTable, &pIter->pIMemTable);
tsdbTakeReadSnap(pTsdb, &pIter->pReadSnap);
STbData *pMem = NULL;
if (pIter->pMemTable) {
tsdbGetTbDataFromMemTable(pIter->pMemTable, suid, uid, &pMem);
if (pIter->pReadSnap->pMem) {
tsdbGetTbDataFromMemTable(pIter->pReadSnap->pMem, suid, uid, &pMem);
}
STbData *pIMem = NULL;
if (pIter->pIMemTable) {
tsdbGetTbDataFromMemTable(pIter->pIMemTable, suid, uid, &pIMem);
if (pIter->pReadSnap->pIMem) {
tsdbGetTbDataFromMemTable(pIter->pReadSnap->pIMem, suid, uid, &pIMem);
}
pIter->pTsdb = pTsdb;
@ -821,7 +822,7 @@ static int32_t nextRowIterOpen(CacheNextRowIter *pIter, tb_uid_t uid, STsdb *pTs
SDelIdx delIdx;
SDelFile *pDelFile = tsdbFSStateGetDelFile(pTsdb->pFS->cState);
SDelFile *pDelFile = pIter->pReadSnap->fs.pDelFile;
if (pDelFile) {
SDelFReader *pDelFReader;
@ -846,6 +847,7 @@ static int32_t nextRowIterOpen(CacheNextRowIter *pIter, tb_uid_t uid, STsdb *pTs
pIter->fsState.state = SFSNEXTROW_FS;
pIter->fsState.pTsdb = pTsdb;
pIter->fsState.aDFileSet = pIter->pReadSnap->fs.aDFileSet;
pIter->fsState.pBlockIdxExp = &pIter->idx;
pIter->input[0] = (TsdbNextRowState){&pIter->memRow, true, false, &pIter->memState, getNextRowFromMem, NULL};
@ -885,7 +887,7 @@ static int32_t nextRowIterClose(CacheNextRowIter *pIter) {
taosArrayDestroy(pIter->pSkyline);
}
tsdbUntakeMemSnapshot(pIter->pTsdb, pIter->pMemTable, pIter->pIMemTable);
tsdbUntakeReadSnap(pIter->pTsdb, pIter->pReadSnap);
return code;
_err:
@ -1172,480 +1174,480 @@ _err:
return code;
}
static int32_t mergeLastRow(tb_uid_t uid, STsdb *pTsdb, bool *dup, STSRow **ppRow) {
int32_t code = 0;
SArray *pSkyline = NULL;
// static int32_t mergeLastRow(tb_uid_t uid, STsdb *pTsdb, bool *dup, STSRow **ppRow) {
// int32_t code = 0;
// SArray *pSkyline = NULL;
STSchema *pTSchema = metaGetTbTSchema(pTsdb->pVnode->pMeta, uid, -1);
int16_t nCol = pTSchema->numOfCols;
SArray *pColArray = taosArrayInit(nCol, sizeof(SColVal));
// STSchema *pTSchema = metaGetTbTSchema(pTsdb->pVnode->pMeta, uid, -1);
// int16_t nCol = pTSchema->numOfCols;
// SArray *pColArray = taosArrayInit(nCol, sizeof(SColVal));
tb_uid_t suid = getTableSuidByUid(uid, pTsdb);
// tb_uid_t suid = getTableSuidByUid(uid, pTsdb);
STbData *pMem = NULL;
if (pTsdb->mem) {
tsdbGetTbDataFromMemTable(pTsdb->mem, suid, uid, &pMem);
}
// STbData *pMem = NULL;
// if (pTsdb->mem) {
// tsdbGetTbDataFromMemTable(pTsdb->mem, suid, uid, &pMem);
// }
STbData *pIMem = NULL;
if (pTsdb->imem) {
tsdbGetTbDataFromMemTable(pTsdb->imem, suid, uid, &pIMem);
}
// STbData *pIMem = NULL;
// if (pTsdb->imem) {
// tsdbGetTbDataFromMemTable(pTsdb->imem, suid, uid, &pIMem);
// }
*ppRow = NULL;
// *ppRow = NULL;
pSkyline = taosArrayInit(32, sizeof(TSDBKEY));
// pSkyline = taosArrayInit(32, sizeof(TSDBKEY));
SDelIdx delIdx;
// SDelIdx delIdx;
SDelFile *pDelFile = tsdbFSStateGetDelFile(pTsdb->pFS->cState);
if (pDelFile) {
SDelFReader *pDelFReader;
// SDelFile *pDelFile = tsdbFSStateGetDelFile(pTsdb->pFS->cState);
// if (pDelFile) {
// SDelFReader *pDelFReader;
code = tsdbDelFReaderOpen(&pDelFReader, pDelFile, pTsdb, NULL);
if (code) goto _err;
// code = tsdbDelFReaderOpen(&pDelFReader, pDelFile, pTsdb, NULL);
// if (code) goto _err;
code = getTableDelIdx(pDelFReader, suid, uid, &delIdx);
if (code) goto _err;
// code = getTableDelIdx(pDelFReader, suid, uid, &delIdx);
// if (code) goto _err;
code = getTableDelSkyline(pMem, pIMem, pDelFReader, &delIdx, pSkyline);
if (code) goto _err;
// code = getTableDelSkyline(pMem, pIMem, pDelFReader, &delIdx, pSkyline);
// if (code) goto _err;
tsdbDelFReaderClose(&pDelFReader);
} else {
code = getTableDelSkyline(pMem, pIMem, NULL, NULL, pSkyline);
if (code) goto _err;
}
// tsdbDelFReaderClose(&pDelFReader);
// } else {
// code = getTableDelSkyline(pMem, pIMem, NULL, NULL, pSkyline);
// if (code) goto _err;
// }
int64_t iSkyline = taosArrayGetSize(pSkyline) - 1;
// int64_t iSkyline = taosArrayGetSize(pSkyline) - 1;
SBlockIdx idx = {.suid = suid, .uid = uid};
// SBlockIdx idx = {.suid = suid, .uid = uid};
SFSNextRowIter fsState = {0};
fsState.state = SFSNEXTROW_FS;
fsState.pTsdb = pTsdb;
fsState.pBlockIdxExp = &idx;
// SFSNextRowIter fsState = {0};
// fsState.state = SFSNEXTROW_FS;
// fsState.pTsdb = pTsdb;
// fsState.pBlockIdxExp = &idx;
SMemNextRowIter memState = {0};
SMemNextRowIter imemState = {0};
TSDBROW memRow, imemRow, fsRow;
// SMemNextRowIter memState = {0};
// SMemNextRowIter imemState = {0};
// TSDBROW memRow, imemRow, fsRow;
TsdbNextRowState input[3] = {{&memRow, true, false, &memState, getNextRowFromMem, NULL},
{&imemRow, true, false, &imemState, getNextRowFromMem, NULL},
{&fsRow, false, true, &fsState, getNextRowFromFS, clearNextRowFromFS}};
// TsdbNextRowState input[3] = {{&memRow, true, false, &memState, getNextRowFromMem, NULL},
// {&imemRow, true, false, &imemState, getNextRowFromMem, NULL},
// {&fsRow, false, true, &fsState, getNextRowFromFS, clearNextRowFromFS}};
if (pMem) {
memState.pMem = pMem;
memState.state = SMEMNEXTROW_ENTER;
input[0].stop = false;
input[0].next = true;
}
if (pIMem) {
imemState.pMem = pIMem;
imemState.state = SMEMNEXTROW_ENTER;
input[1].stop = false;
input[1].next = true;
}
// if (pMem) {
// memState.pMem = pMem;
// memState.state = SMEMNEXTROW_ENTER;
// input[0].stop = false;
// input[0].next = true;
// }
// if (pIMem) {
// imemState.pMem = pIMem;
// imemState.state = SMEMNEXTROW_ENTER;
// input[1].stop = false;
// input[1].next = true;
// }
int16_t nilColCount = nCol - 1; // count of null & none cols
int iCol = 0; // index of first nil col index from left to right
bool setICol = false;
// int16_t nilColCount = nCol - 1; // count of null & none cols
// int iCol = 0; // index of first nil col index from left to right
// bool setICol = false;
do {
for (int i = 0; i < 3; ++i) {
if (input[i].next && !input[i].stop) {
if (input[i].pRow == NULL) {
code = input[i].nextRowFn(input[i].iter, &input[i].pRow);
if (code) goto _err;
// do {
// for (int i = 0; i < 3; ++i) {
// if (input[i].next && !input[i].stop) {
// if (input[i].pRow == NULL) {
// code = input[i].nextRowFn(input[i].iter, &input[i].pRow);
// if (code) goto _err;
if (input[i].pRow == NULL) {
input[i].stop = true;
input[i].next = false;
}
}
}
}
// if (input[i].pRow == NULL) {
// input[i].stop = true;
// input[i].next = false;
// }
// }
// }
// }
if (input[0].stop && input[1].stop && input[2].stop) {
break;
}
// if (input[0].stop && input[1].stop && input[2].stop) {
// break;
// }
// select maxpoint(s) from mem, imem, fs
TSDBROW *max[3] = {0};
int iMax[3] = {-1, -1, -1};
int nMax = 0;
TSKEY maxKey = TSKEY_MIN;
// // select maxpoint(s) from mem, imem, fs
// TSDBROW *max[3] = {0};
// int iMax[3] = {-1, -1, -1};
// int nMax = 0;
// TSKEY maxKey = TSKEY_MIN;
for (int i = 0; i < 3; ++i) {
if (!input[i].stop && input[i].pRow != NULL) {
TSDBKEY key = TSDBROW_KEY(input[i].pRow);
// for (int i = 0; i < 3; ++i) {
// if (!input[i].stop && input[i].pRow != NULL) {
// TSDBKEY key = TSDBROW_KEY(input[i].pRow);
// merging & deduplicating on client side
if (maxKey <= key.ts) {
if (maxKey < key.ts) {
nMax = 0;
maxKey = key.ts;
}
// // merging & deduplicating on client side
// if (maxKey <= key.ts) {
// if (maxKey < key.ts) {
// nMax = 0;
// maxKey = key.ts;
// }
iMax[nMax] = i;
max[nMax++] = input[i].pRow;
}
}
}
// iMax[nMax] = i;
// max[nMax++] = input[i].pRow;
// }
// }
// }
// delete detection
TSDBROW *merge[3] = {0};
int iMerge[3] = {-1, -1, -1};
int nMerge = 0;
for (int i = 0; i < nMax; ++i) {
TSDBKEY maxKey = TSDBROW_KEY(max[i]);
// // delete detection
// TSDBROW *merge[3] = {0};
// int iMerge[3] = {-1, -1, -1};
// int nMerge = 0;
// for (int i = 0; i < nMax; ++i) {
// TSDBKEY maxKey = TSDBROW_KEY(max[i]);
bool deleted = tsdbKeyDeleted(&maxKey, pSkyline, &iSkyline);
if (!deleted) {
iMerge[nMerge] = i;
merge[nMerge++] = max[i];
}
// bool deleted = tsdbKeyDeleted(&maxKey, pSkyline, &iSkyline);
// if (!deleted) {
// iMerge[nMerge] = i;
// merge[nMerge++] = max[i];
// }
input[iMax[i]].next = deleted;
}
// input[iMax[i]].next = deleted;
// }
// merge if nMerge > 1
if (nMerge > 0) {
*dup = false;
// // merge if nMerge > 1
// if (nMerge > 0) {
// *dup = false;
if (nMerge == 1) {
code = tsRowFromTsdbRow(pTSchema, merge[nMerge - 1], ppRow);
if (code) goto _err;
} else {
// merge 2 or 3 rows
SRowMerger merger = {0};
// if (nMerge == 1) {
// code = tsRowFromTsdbRow(pTSchema, merge[nMerge - 1], ppRow);
// if (code) goto _err;
// } else {
// // merge 2 or 3 rows
// SRowMerger merger = {0};
tRowMergerInit(&merger, merge[0], pTSchema);
for (int i = 1; i < nMerge; ++i) {
tRowMerge(&merger, merge[i]);
}
tRowMergerGetRow(&merger, ppRow);
tRowMergerClear(&merger);
}
}
// tRowMergerInit(&merger, merge[0], pTSchema);
// for (int i = 1; i < nMerge; ++i) {
// tRowMerge(&merger, merge[i]);
// }
// tRowMergerGetRow(&merger, ppRow);
// tRowMergerClear(&merger);
// }
// }
} while (1);
// } while (1);
for (int i = 0; i < 3; ++i) {
if (input[i].nextRowClearFn) {
input[i].nextRowClearFn(input[i].iter);
}
}
if (pSkyline) {
taosArrayDestroy(pSkyline);
}
taosMemoryFreeClear(pTSchema);
// for (int i = 0; i < 3; ++i) {
// if (input[i].nextRowClearFn) {
// input[i].nextRowClearFn(input[i].iter);
// }
// }
// if (pSkyline) {
// taosArrayDestroy(pSkyline);
// }
// taosMemoryFreeClear(pTSchema);
return code;
_err:
for (int i = 0; i < 3; ++i) {
if (input[i].nextRowClearFn) {
input[i].nextRowClearFn(input[i].iter);
}
}
if (pSkyline) {
taosArrayDestroy(pSkyline);
}
taosMemoryFreeClear(pTSchema);
tsdbError("vgId:%d merge last_row failed since %s", TD_VID(pTsdb->pVnode), tstrerror(code));
return code;
}
// return code;
// _err:
// for (int i = 0; i < 3; ++i) {
// if (input[i].nextRowClearFn) {
// input[i].nextRowClearFn(input[i].iter);
// }
// }
// if (pSkyline) {
// taosArrayDestroy(pSkyline);
// }
// taosMemoryFreeClear(pTSchema);
// tsdbError("vgId:%d merge last_row failed since %s", TD_VID(pTsdb->pVnode), tstrerror(code));
// return code;
// }
// static int32_t mergeLast(tb_uid_t uid, STsdb *pTsdb, STSRow **ppRow) {
static int32_t mergeLast(tb_uid_t uid, STsdb *pTsdb, SArray **ppLastArray) {
int32_t code = 0;
SArray *pSkyline = NULL;
STSRow *pRow = NULL;
STSRow **ppRow = &pRow;
// static int32_t mergeLast(tb_uid_t uid, STsdb *pTsdb, SArray **ppLastArray) {
// int32_t code = 0;
// SArray *pSkyline = NULL;
// STSRow *pRow = NULL;
// STSRow **ppRow = &pRow;
STSchema *pTSchema = metaGetTbTSchema(pTsdb->pVnode->pMeta, uid, -1);
int16_t nCol = pTSchema->numOfCols;
// SArray *pColArray = taosArrayInit(nCol, sizeof(SColVal));
SArray *pColArray = taosArrayInit(nCol, sizeof(SLastCol));
// STSchema *pTSchema = metaGetTbTSchema(pTsdb->pVnode->pMeta, uid, -1);
// int16_t nCol = pTSchema->numOfCols;
// // SArray *pColArray = taosArrayInit(nCol, sizeof(SColVal));
// SArray *pColArray = taosArrayInit(nCol, sizeof(SLastCol));
tb_uid_t suid = getTableSuidByUid(uid, pTsdb);
// tb_uid_t suid = getTableSuidByUid(uid, pTsdb);
STbData *pMem = NULL;
if (pTsdb->mem) {
tsdbGetTbDataFromMemTable(pTsdb->mem, suid, uid, &pMem);
}
// STbData *pMem = NULL;
// if (pTsdb->mem) {
// tsdbGetTbDataFromMemTable(pTsdb->mem, suid, uid, &pMem);
// }
STbData *pIMem = NULL;
if (pTsdb->imem) {
tsdbGetTbDataFromMemTable(pTsdb->imem, suid, uid, &pIMem);
}
// STbData *pIMem = NULL;
// if (pTsdb->imem) {
// tsdbGetTbDataFromMemTable(pTsdb->imem, suid, uid, &pIMem);
// }
*ppLastArray = NULL;
// *ppLastArray = NULL;
pSkyline = taosArrayInit(32, sizeof(TSDBKEY));
// pSkyline = taosArrayInit(32, sizeof(TSDBKEY));
SDelIdx delIdx;
// SDelIdx delIdx;
SDelFile *pDelFile = tsdbFSStateGetDelFile(pTsdb->pFS->cState);
if (pDelFile) {
SDelFReader *pDelFReader;
// SDelFile *pDelFile = tsdbFSStateGetDelFile(pTsdb->pFS->cState);
// if (pDelFile) {
// SDelFReader *pDelFReader;
code = tsdbDelFReaderOpen(&pDelFReader, pDelFile, pTsdb, NULL);
if (code) goto _err;
// code = tsdbDelFReaderOpen(&pDelFReader, pDelFile, pTsdb, NULL);
// if (code) goto _err;
code = getTableDelIdx(pDelFReader, suid, uid, &delIdx);
if (code) goto _err;
// code = getTableDelIdx(pDelFReader, suid, uid, &delIdx);
// if (code) goto _err;
code = getTableDelSkyline(pMem, pIMem, pDelFReader, &delIdx, pSkyline);
if (code) goto _err;
// code = getTableDelSkyline(pMem, pIMem, pDelFReader, &delIdx, pSkyline);
// if (code) goto _err;
tsdbDelFReaderClose(&pDelFReader);
} else {
code = getTableDelSkyline(pMem, pIMem, NULL, NULL, pSkyline);
if (code) goto _err;
}
// tsdbDelFReaderClose(&pDelFReader);
// } else {
// code = getTableDelSkyline(pMem, pIMem, NULL, NULL, pSkyline);
// if (code) goto _err;
// }
int64_t iSkyline = taosArrayGetSize(pSkyline) - 1;
// int64_t iSkyline = taosArrayGetSize(pSkyline) - 1;
SBlockIdx idx = {.suid = suid, .uid = uid};
// SBlockIdx idx = {.suid = suid, .uid = uid};
SFSNextRowIter fsState = {0};
fsState.state = SFSNEXTROW_FS;
fsState.pTsdb = pTsdb;
fsState.pBlockIdxExp = &idx;
// SFSNextRowIter fsState = {0};
// fsState.state = SFSNEXTROW_FS;
// fsState.pTsdb = pTsdb;
// fsState.pBlockIdxExp = &idx;
SMemNextRowIter memState = {0};
SMemNextRowIter imemState = {0};
TSDBROW memRow, imemRow, fsRow;
// SMemNextRowIter memState = {0};
// SMemNextRowIter imemState = {0};
// TSDBROW memRow, imemRow, fsRow;
TsdbNextRowState input[3] = {{&memRow, true, false, &memState, getNextRowFromMem, NULL},
{&imemRow, true, false, &imemState, getNextRowFromMem, NULL},
{&fsRow, false, true, &fsState, getNextRowFromFS, clearNextRowFromFS}};
// TsdbNextRowState input[3] = {{&memRow, true, false, &memState, getNextRowFromMem, NULL},
// {&imemRow, true, false, &imemState, getNextRowFromMem, NULL},
// {&fsRow, false, true, &fsState, getNextRowFromFS, clearNextRowFromFS}};
if (pMem) {
memState.pMem = pMem;
memState.state = SMEMNEXTROW_ENTER;
input[0].stop = false;
input[0].next = true;
}
if (pIMem) {
imemState.pMem = pIMem;
imemState.state = SMEMNEXTROW_ENTER;
input[1].stop = false;
input[1].next = true;
}
// if (pMem) {
// memState.pMem = pMem;
// memState.state = SMEMNEXTROW_ENTER;
// input[0].stop = false;
// input[0].next = true;
// }
// if (pIMem) {
// imemState.pMem = pIMem;
// imemState.state = SMEMNEXTROW_ENTER;
// input[1].stop = false;
// input[1].next = true;
// }
int16_t nilColCount = nCol - 1; // count of null & none cols
int iCol = 0; // index of first nil col index from left to right
bool setICol = false;
// int16_t nilColCount = nCol - 1; // count of null & none cols
// int iCol = 0; // index of first nil col index from left to right
// bool setICol = false;
do {
for (int i = 0; i < 3; ++i) {
if (input[i].next && !input[i].stop) {
code = input[i].nextRowFn(input[i].iter, &input[i].pRow);
if (code) goto _err;
// do {
// for (int i = 0; i < 3; ++i) {
// if (input[i].next && !input[i].stop) {
// code = input[i].nextRowFn(input[i].iter, &input[i].pRow);
// if (code) goto _err;
if (input[i].pRow == NULL) {
input[i].stop = true;
input[i].next = false;
}
}
}
// if (input[i].pRow == NULL) {
// input[i].stop = true;
// input[i].next = false;
// }
// }
// }
if (input[0].stop && input[1].stop && input[2].stop) {
break;
}
// if (input[0].stop && input[1].stop && input[2].stop) {
// break;
// }
// select maxpoint(s) from mem, imem, fs
TSDBROW *max[3] = {0};
int iMax[3] = {-1, -1, -1};
int nMax = 0;
TSKEY maxKey = TSKEY_MIN;
// // select maxpoint(s) from mem, imem, fs
// TSDBROW *max[3] = {0};
// int iMax[3] = {-1, -1, -1};
// int nMax = 0;
// TSKEY maxKey = TSKEY_MIN;
for (int i = 0; i < 3; ++i) {
if (!input[i].stop && input[i].pRow != NULL) {
TSDBKEY key = TSDBROW_KEY(input[i].pRow);
// for (int i = 0; i < 3; ++i) {
// if (!input[i].stop && input[i].pRow != NULL) {
// TSDBKEY key = TSDBROW_KEY(input[i].pRow);
// merging & deduplicating on client side
if (maxKey <= key.ts) {
if (maxKey < key.ts) {
nMax = 0;
maxKey = key.ts;
}
// // merging & deduplicating on client side
// if (maxKey <= key.ts) {
// if (maxKey < key.ts) {
// nMax = 0;
// maxKey = key.ts;
// }
iMax[nMax] = i;
max[nMax++] = input[i].pRow;
}
}
}
// iMax[nMax] = i;
// max[nMax++] = input[i].pRow;
// }
// }
// }
// delete detection
TSDBROW *merge[3] = {0};
int iMerge[3] = {-1, -1, -1};
int nMerge = 0;
for (int i = 0; i < nMax; ++i) {
TSDBKEY maxKey = TSDBROW_KEY(max[i]);
// // delete detection
// TSDBROW *merge[3] = {0};
// int iMerge[3] = {-1, -1, -1};
// int nMerge = 0;
// for (int i = 0; i < nMax; ++i) {
// TSDBKEY maxKey = TSDBROW_KEY(max[i]);
bool deleted = tsdbKeyDeleted(&maxKey, pSkyline, &iSkyline);
if (!deleted) {
iMerge[nMerge] = iMax[i];
merge[nMerge++] = max[i];
}
// bool deleted = tsdbKeyDeleted(&maxKey, pSkyline, &iSkyline);
// if (!deleted) {
// iMerge[nMerge] = iMax[i];
// merge[nMerge++] = max[i];
// }
input[iMax[i]].next = deleted;
}
// input[iMax[i]].next = deleted;
// }
// merge if nMerge > 1
if (nMerge > 0) {
if (nMerge == 1) {
code = tsRowFromTsdbRow(pTSchema, merge[nMerge - 1], ppRow);
if (code) goto _err;
} else {
// merge 2 or 3 rows
SRowMerger merger = {0};
// // merge if nMerge > 1
// if (nMerge > 0) {
// if (nMerge == 1) {
// code = tsRowFromTsdbRow(pTSchema, merge[nMerge - 1], ppRow);
// if (code) goto _err;
// } else {
// // merge 2 or 3 rows
// SRowMerger merger = {0};
tRowMergerInit(&merger, merge[0], pTSchema);
for (int i = 1; i < nMerge; ++i) {
tRowMerge(&merger, merge[i]);
}
tRowMergerGetRow(&merger, ppRow);
tRowMergerClear(&merger);
}
} else {
/* *ppRow = NULL; */
/* return code; */
continue;
}
// tRowMergerInit(&merger, merge[0], pTSchema);
// for (int i = 1; i < nMerge; ++i) {
// tRowMerge(&merger, merge[i]);
// }
// tRowMergerGetRow(&merger, ppRow);
// tRowMergerClear(&merger);
// }
// } else {
// /* *ppRow = NULL; */
// /* return code; */
// continue;
// }
if (iCol == 0) {
STColumn *pTColumn = &pTSchema->columns[0];
SColVal *pColVal = &(SColVal){0};
// if (iCol == 0) {
// STColumn *pTColumn = &pTSchema->columns[0];
// SColVal *pColVal = &(SColVal){0};
*pColVal = COL_VAL_VALUE(pTColumn->colId, pTColumn->type, (SValue){.ts = maxKey});
// *pColVal = COL_VAL_VALUE(pTColumn->colId, pTColumn->type, (SValue){.ts = maxKey});
// if (taosArrayPush(pColArray, pColVal) == NULL) {
if (taosArrayPush(pColArray, &(SLastCol){.ts = maxKey, .colVal = *pColVal}) == NULL) {
code = TSDB_CODE_OUT_OF_MEMORY;
goto _err;
}
// // if (taosArrayPush(pColArray, pColVal) == NULL) {
// if (taosArrayPush(pColArray, &(SLastCol){.ts = maxKey, .colVal = *pColVal}) == NULL) {
// code = TSDB_CODE_OUT_OF_MEMORY;
// goto _err;
// }
++iCol;
// ++iCol;
setICol = false;
for (int16_t i = iCol; i < nCol; ++i) {
// tsdbRowGetColVal(*ppRow, pTSchema, i, pColVal);
tTSRowGetVal(*ppRow, pTSchema, i, pColVal);
// if (taosArrayPush(pColArray, pColVal) == NULL) {
if (taosArrayPush(pColArray, &(SLastCol){.ts = maxKey, .colVal = *pColVal}) == NULL) {
code = TSDB_CODE_OUT_OF_MEMORY;
goto _err;
}
// setICol = false;
// for (int16_t i = iCol; i < nCol; ++i) {
// // tsdbRowGetColVal(*ppRow, pTSchema, i, pColVal);
// tTSRowGetVal(*ppRow, pTSchema, i, pColVal);
// // if (taosArrayPush(pColArray, pColVal) == NULL) {
// if (taosArrayPush(pColArray, &(SLastCol){.ts = maxKey, .colVal = *pColVal}) == NULL) {
// code = TSDB_CODE_OUT_OF_MEMORY;
// goto _err;
// }
if (pColVal->isNull || pColVal->isNone) {
for (int j = 0; j < nMerge; ++j) {
SColVal jColVal = {0};
tsdbRowGetColVal(merge[j], pTSchema, i, &jColVal);
if (jColVal.isNull || jColVal.isNone) {
input[iMerge[j]].next = true;
}
}
if (!setICol) {
iCol = i;
setICol = true;
}
} else {
--nilColCount;
}
}
// if (pColVal->isNull || pColVal->isNone) {
// for (int j = 0; j < nMerge; ++j) {
// SColVal jColVal = {0};
// tsdbRowGetColVal(merge[j], pTSchema, i, &jColVal);
// if (jColVal.isNull || jColVal.isNone) {
// input[iMerge[j]].next = true;
// }
// }
// if (!setICol) {
// iCol = i;
// setICol = true;
// }
// } else {
// --nilColCount;
// }
// }
if (*ppRow) {
taosMemoryFreeClear(*ppRow);
}
// if (*ppRow) {
// taosMemoryFreeClear(*ppRow);
// }
continue;
}
// continue;
// }
setICol = false;
for (int16_t i = iCol; i < nCol; ++i) {
SColVal colVal = {0};
tTSRowGetVal(*ppRow, pTSchema, i, &colVal);
TSKEY rowTs = (*ppRow)->ts;
// setICol = false;
// for (int16_t i = iCol; i < nCol; ++i) {
// SColVal colVal = {0};
// tTSRowGetVal(*ppRow, pTSchema, i, &colVal);
// TSKEY rowTs = (*ppRow)->ts;
// SColVal *tColVal = (SColVal *)taosArrayGet(pColArray, i);
SLastCol *tTsVal = (SLastCol *)taosArrayGet(pColArray, i);
SColVal *tColVal = &tTsVal->colVal;
// // SColVal *tColVal = (SColVal *)taosArrayGet(pColArray, i);
// SLastCol *tTsVal = (SLastCol *)taosArrayGet(pColArray, i);
// SColVal *tColVal = &tTsVal->colVal;
if (!colVal.isNone && !colVal.isNull) {
if (tColVal->isNull || tColVal->isNone) {
// taosArraySet(pColArray, i, &colVal);
taosArraySet(pColArray, i, &(SLastCol){.ts = rowTs, .colVal = colVal});
--nilColCount;
}
} else {
if ((tColVal->isNull || tColVal->isNone) && !setICol) {
iCol = i;
setICol = true;
// if (!colVal.isNone && !colVal.isNull) {
// if (tColVal->isNull || tColVal->isNone) {
// // taosArraySet(pColArray, i, &colVal);
// taosArraySet(pColArray, i, &(SLastCol){.ts = rowTs, .colVal = colVal});
// --nilColCount;
// }
// } else {
// if ((tColVal->isNull || tColVal->isNone) && !setICol) {
// iCol = i;
// setICol = true;
for (int j = 0; j < nMerge; ++j) {
SColVal jColVal = {0};
tsdbRowGetColVal(merge[j], pTSchema, i, &jColVal);
if (jColVal.isNull || jColVal.isNone) {
input[iMerge[j]].next = true;
}
}
}
}
}
// for (int j = 0; j < nMerge; ++j) {
// SColVal jColVal = {0};
// tsdbRowGetColVal(merge[j], pTSchema, i, &jColVal);
// if (jColVal.isNull || jColVal.isNone) {
// input[iMerge[j]].next = true;
// }
// }
// }
// }
// }
if (*ppRow) {
taosMemoryFreeClear(*ppRow);
}
} while (nilColCount > 0);
// if (*ppRow) {
// taosMemoryFreeClear(*ppRow);
// }
// } while (nilColCount > 0);
// if () new ts row from pColArray if non empty
/* if (taosArrayGetSize(pColArray) == nCol) { */
/* code = tdSTSRowNew(pColArray, pTSchema, ppRow); */
/* if (code) goto _err; */
/* } */
/* taosArrayDestroy(pColArray); */
if (taosArrayGetSize(pColArray) <= 0) {
*ppLastArray = NULL;
taosArrayDestroy(pColArray);
} else {
*ppLastArray = pColArray;
}
if (*ppRow) {
taosMemoryFreeClear(*ppRow);
}
// // if () new ts row from pColArray if non empty
// /* if (taosArrayGetSize(pColArray) == nCol) { */
// /* code = tdSTSRowNew(pColArray, pTSchema, ppRow); */
// /* if (code) goto _err; */
// /* } */
// /* taosArrayDestroy(pColArray); */
// if (taosArrayGetSize(pColArray) <= 0) {
// *ppLastArray = NULL;
// taosArrayDestroy(pColArray);
// } else {
// *ppLastArray = pColArray;
// }
// if (*ppRow) {
// taosMemoryFreeClear(*ppRow);
// }
for (int i = 0; i < 3; ++i) {
if (input[i].nextRowClearFn) {
input[i].nextRowClearFn(input[i].iter);
}
}
if (pSkyline) {
taosArrayDestroy(pSkyline);
}
taosMemoryFreeClear(pTSchema);
// for (int i = 0; i < 3; ++i) {
// if (input[i].nextRowClearFn) {
// input[i].nextRowClearFn(input[i].iter);
// }
// }
// if (pSkyline) {
// taosArrayDestroy(pSkyline);
// }
// taosMemoryFreeClear(pTSchema);
return code;
_err:
taosArrayDestroy(pColArray);
if (*ppRow) {
taosMemoryFreeClear(*ppRow);
}
for (int i = 0; i < 3; ++i) {
if (input[i].nextRowClearFn) {
input[i].nextRowClearFn(input[i].iter);
}
}
if (pSkyline) {
taosArrayDestroy(pSkyline);
}
taosMemoryFreeClear(pTSchema);
tsdbError("vgId:%d merge last_row failed since %s", TD_VID(pTsdb->pVnode), tstrerror(code));
return code;
}
// return code;
// _err:
// taosArrayDestroy(pColArray);
// if (*ppRow) {
// taosMemoryFreeClear(*ppRow);
// }
// for (int i = 0; i < 3; ++i) {
// if (input[i].nextRowClearFn) {
// input[i].nextRowClearFn(input[i].iter);
// }
// }
// if (pSkyline) {
// taosArrayDestroy(pSkyline);
// }
// taosMemoryFreeClear(pTSchema);
// tsdbError("vgId:%d merge last_row failed since %s", TD_VID(pTsdb->pVnode), tstrerror(code));
// return code;
// }
int32_t tsdbCacheGetLastrowH(SLRUCache *pCache, tb_uid_t uid, STsdb *pTsdb, LRUHandle **handle) {
int32_t code = 0;

73
source/dnode/vnode/src/tsdb/tsdbCommit.c
View File

@ -29,6 +29,7 @@ typedef struct {
int32_t minRow;
int32_t maxRow;
int8_t cmprAlg;
STsdbFS fs;
// --------------
TSKEY nextKey; // reset by each table commit
int32_t commitFid;
@ -119,9 +120,6 @@ int32_t tsdbCommit(STsdb *pTsdb) {
code = tsdbCommitDel(&commith);
if (code) goto _err;
code = tsdbCommitCache(&commith);
if (code) goto _err;
// end commit
code = tsdbEndCommit(&commith, 0);
if (code) goto _err;
@ -158,7 +156,7 @@ static int32_t tsdbCommitDelStart(SCommitter *pCommitter) {
goto _err;
}
SDelFile *pDelFileR = pTsdb->pFS->nState->pDelFile;
SDelFile *pDelFileR = pCommitter->fs.pDelFile;
if (pDelFileR) {
code = tsdbDelFReaderOpen(&pCommitter->pDelFReader, pDelFileR, pTsdb, NULL);
if (code) goto _err;
@ -247,7 +245,7 @@ static int32_t tsdbCommitDelEnd(SCommitter *pCommitter) {
code = tsdbUpdateDelFileHdr(pCommitter->pDelFWriter);
if (code) goto _err;
code = tsdbFSStateUpsertDelFile(pTsdb->pFS->nState, &pCommitter->pDelFWriter->fDel);
code = tsdbFSUpsertDelFile(&pCommitter->fs, &pCommitter->pDelFWriter->fDel);
if (code) goto _err;
code = tsdbDelFWriterClose(&pCommitter->pDelFWriter, 1);
@ -273,7 +271,6 @@ static int32_t tsdbCommitFileDataStart(SCommitter *pCommitter) {
int32_t code = 0;
STsdb *pTsdb = pCommitter->pTsdb;
SDFileSet *pRSet = NULL;
SDFileSet wSet;
// memory
pCommitter->nextKey = TSKEY_MAX;
@ -282,7 +279,8 @@ static int32_t tsdbCommitFileDataStart(SCommitter *pCommitter) {
taosArrayClear(pCommitter->aBlockIdx);
tMapDataReset(&pCommitter->oBlockMap);
tBlockDataReset(&pCommitter->oBlockData);
pRSet = tsdbFSStateGetDFileSet(pTsdb->pFS->nState, pCommitter->commitFid, TD_EQ);
pRSet = (SDFileSet *)taosArraySearch(pCommitter->fs.aDFileSet, &(SDFileSet){.fid = pCommitter->commitFid},
tDFileSetCmprFn, TD_EQ);
if (pRSet) {
code = tsdbDataFReaderOpen(&pCommitter->pReader, pTsdb, pRSet);
if (code) goto _err;
@ -292,23 +290,29 @@ static int32_t tsdbCommitFileDataStart(SCommitter *pCommitter) {
}
// new
SHeadFile fHead;
SDataFile fData;
SLastFile fLast;
SSmaFile fSma;
SDFileSet wSet = {.pHeadF = &fHead, .pDataF = &fData, .pLastF = &fLast, .pSmaF = &fSma};
taosArrayClear(pCommitter->aBlockIdxN);
tMapDataReset(&pCommitter->nBlockMap);
tBlockDataReset(&pCommitter->nBlockData);
if (pRSet) {
wSet = (SDFileSet){.diskId = pRSet->diskId,
.fid = pCommitter->commitFid,
.fHead = {.commitID = pCommitter->commitID, .offset = 0, .size = 0},
.fData = pRSet->fData,
.fLast = {.commitID = pCommitter->commitID, .size = 0},
.fSma = pRSet->fSma};
wSet.diskId = pRSet->diskId;
wSet.fid = pCommitter->commitFid;
fHead = (SHeadFile){.commitID = pCommitter->commitID, .offset = 0, .size = 0};
fData = *pRSet->pDataF;
fLast = (SLastFile){.commitID = pCommitter->commitID, .size = 0};
fSma = *pRSet->pSmaF;
} else {
wSet = (SDFileSet){.diskId = (SDiskID){.level = 0, .id = 0},
.fid = pCommitter->commitFid,
.fHead = {.commitID = pCommitter->commitID, .offset = 0, .size = 0},
.fData = {.commitID = pCommitter->commitID, .size = 0},
.fLast = {.commitID = pCommitter->commitID, .size = 0},
.fSma = {.commitID = pCommitter->commitID, .size = 0}};
wSet.diskId = (SDiskID){.level = 0, .id = 0};
wSet.fid = pCommitter->commitFid;
fHead = (SHeadFile){.commitID = pCommitter->commitID, .offset = 0, .size = 0};
fData = (SDataFile){.commitID = pCommitter->commitID, .size = 0};
fLast = (SLastFile){.commitID = pCommitter->commitID, .size = 0};
fSma = (SSmaFile){.commitID = pCommitter->commitID, .size = 0};
}
code = tsdbDataFWriterOpen(&pCommitter->pWriter, pTsdb, &wSet);
if (code) goto _err;
@ -855,7 +859,7 @@ static int32_t tsdbCommitFileDataEnd(SCommitter *pCommitter) {
if (code) goto _err;
// upsert SDFileSet
code = tsdbFSStateUpsertDFileSet(pCommitter->pTsdb->pFS->nState, tsdbDataFWriterGetWSet(pCommitter->pWriter));
code = tsdbFSUpsertFSet(&pCommitter->fs, &pCommitter->pWriter->wSet);
if (code) goto _err;
// close and sync
@ -973,7 +977,7 @@ static int32_t tsdbStartCommit(STsdb *pTsdb, SCommitter *pCommitter) {
pCommitter->maxRow = pTsdb->pVnode->config.tsdbCfg.maxRows;
pCommitter->cmprAlg = pTsdb->pVnode->config.tsdbCfg.compression;
code = tsdbFSBegin(pTsdb->pFS);
code = tsdbFSCopy(pTsdb, &pCommitter->fs);
if (code) goto _err;
return code;
@ -1142,28 +1146,33 @@ _err:
return code;
}
static int32_t tsdbCommitCache(SCommitter *pCommitter) {
int32_t code = 0;
// TODO
return code;
}
static int32_t tsdbEndCommit(SCommitter *pCommitter, int32_t eno) {
int32_t code = 0;
STsdb *pTsdb = pCommitter->pTsdb;
SMemTable *pMemTable = pTsdb->imem;
if (eno == 0) {
code = tsdbFSCommit(pTsdb->pFS);
} else {
code = tsdbFSRollback(pTsdb->pFS);
ASSERT(eno == 0);
code = tsdbFSCommit1(pTsdb, &pCommitter->fs);
if (code) goto _err;
// lock
taosThreadRwlockWrlock(&pTsdb->rwLock);
// commit or rollback
code = tsdbFSCommit2(pTsdb, &pCommitter->fs);
if (code) {
taosThreadRwlockUnlock(&pTsdb->rwLock);
goto _err;
}
taosThreadRwlockWrlock(&pTsdb->rwLock);
pTsdb->imem = NULL;
// unlock
taosThreadRwlockUnlock(&pTsdb->rwLock);
tsdbUnrefMemTable(pMemTable);
tsdbFSDestroy(&pCommitter->fs);
tsdbInfo("vgId:%d tsdb end commit", TD_VID(pTsdb->pVnode));
return code;

1372
source/dnode/vnode/src/tsdb/tsdbFS.c
View File

File diff suppressed because it is too large Load Diff

190
source/dnode/vnode/src/tsdb/tsdbFile.c
View File

@ -15,7 +15,7 @@
#include "tsdb.h"
static int32_t tPutHeadFile(uint8_t *p, SHeadFile *pHeadFile) {
int32_t tPutHeadFile(uint8_t *p, SHeadFile *pHeadFile) {
int32_t n = 0;
n += tPutI64v(p ? p + n : p, pHeadFile->commitID);
@ -35,7 +35,7 @@ static int32_t tGetHeadFile(uint8_t *p, SHeadFile *pHeadFile) {
return n;
}
static int32_t tPutDataFile(uint8_t *p, SDataFile *pDataFile) {
int32_t tPutDataFile(uint8_t *p, SDataFile *pDataFile) {
int32_t n = 0;
n += tPutI64v(p ? p + n : p, pDataFile->commitID);
@ -53,7 +53,7 @@ static int32_t tGetDataFile(uint8_t *p, SDataFile *pDataFile) {
return n;
}
static int32_t tPutLastFile(uint8_t *p, SLastFile *pLastFile) {
int32_t tPutLastFile(uint8_t *p, SLastFile *pLastFile) {
int32_t n = 0;
n += tPutI64v(p ? p + n : p, pLastFile->commitID);
@ -71,7 +71,7 @@ static int32_t tGetLastFile(uint8_t *p, SLastFile *pLastFile) {
return n;
}
static int32_t tPutSmaFile(uint8_t *p, SSmaFile *pSmaFile) {
int32_t tPutSmaFile(uint8_t *p, SSmaFile *pSmaFile) {
int32_t n = 0;
n += tPutI64v(p ? p + n : p, pSmaFile->commitID);
@ -90,90 +90,53 @@ static int32_t tGetSmaFile(uint8_t *p, SSmaFile *pSmaFile) {
}
// EXPOSED APIS ==================================================
void tsdbDataFileName(STsdb *pTsdb, SDFileSet *pDFileSet, EDataFileT ftype, char fname[]) {
STfs *pTfs = pTsdb->pVnode->pTfs;
switch (ftype) {
case TSDB_HEAD_FILE:
snprintf(fname, TSDB_FILENAME_LEN - 1, "%s%s%s%sv%df%dver%" PRId64 "%s", tfsGetDiskPath(pTfs, pDFileSet->diskId),
TD_DIRSEP, pTsdb->path, TD_DIRSEP, TD_VID(pTsdb->pVnode), pDFileSet->fid, pDFileSet->fHead.commitID,
".head");
break;
case TSDB_DATA_FILE:
snprintf(fname, TSDB_FILENAME_LEN - 1, "%s%s%s%sv%df%dver%" PRId64 "%s", tfsGetDiskPath(pTfs, pDFileSet->diskId),
TD_DIRSEP, pTsdb->path, TD_DIRSEP, TD_VID(pTsdb->pVnode), pDFileSet->fid, pDFileSet->fData.commitID,
".data");
break;
case TSDB_LAST_FILE:
snprintf(fname, TSDB_FILENAME_LEN - 1, "%s%s%s%sv%df%dver%" PRId64 "%s", tfsGetDiskPath(pTfs, pDFileSet->diskId),
TD_DIRSEP, pTsdb->path, TD_DIRSEP, TD_VID(pTsdb->pVnode), pDFileSet->fid, pDFileSet->fLast.commitID,
".last");
break;
case TSDB_SMA_FILE:
snprintf(fname, TSDB_FILENAME_LEN - 1, "%s%s%s%sv%df%dver%" PRId64 "%s", tfsGetDiskPath(pTfs, pDFileSet->diskId),
TD_DIRSEP, pTsdb->path, TD_DIRSEP, TD_VID(pTsdb->pVnode), pDFileSet->fid, pDFileSet->fSma.commitID,
".sma");
break;
default:
ASSERT(0);
break;
}
void tsdbHeadFileName(STsdb *pTsdb, SDiskID did, int32_t fid, SHeadFile *pHeadF, char fname[]) {
snprintf(fname, TSDB_FILENAME_LEN - 1, "%s%s%s%sv%df%dver%" PRId64 "%s", tfsGetDiskPath(pTsdb->pVnode->pTfs, did),
TD_DIRSEP, pTsdb->path, TD_DIRSEP, TD_VID(pTsdb->pVnode), fid, pHeadF->commitID, ".head");
}
bool tsdbFileIsSame(SDFileSet *pDFileSet1, SDFileSet *pDFileSet2, EDataFileT ftype) {
if (pDFileSet1->diskId.level != pDFileSet2->diskId.level || pDFileSet1->diskId.id != pDFileSet2->diskId.id) {
return false;
}
void tsdbDataFileName(STsdb *pTsdb, SDiskID did, int32_t fid, SDataFile *pDataF, char fname[]) {
snprintf(fname, TSDB_FILENAME_LEN - 1, "%s%s%s%sv%df%dver%" PRId64 "%s", tfsGetDiskPath(pTsdb->pVnode->pTfs, did),
TD_DIRSEP, pTsdb->path, TD_DIRSEP, TD_VID(pTsdb->pVnode), fid, pDataF->commitID, ".data");
}
switch (ftype) {
case TSDB_HEAD_FILE:
return pDFileSet1->fHead.commitID == pDFileSet2->fHead.commitID;
case TSDB_DATA_FILE:
return pDFileSet1->fData.commitID == pDFileSet2->fData.commitID;
case TSDB_LAST_FILE:
return pDFileSet1->fLast.commitID == pDFileSet2->fLast.commitID;
case TSDB_SMA_FILE:
return pDFileSet1->fSma.commitID == pDFileSet2->fSma.commitID;
default:
ASSERT(0);
break;
}
void tsdbLastFileName(STsdb *pTsdb, SDiskID did, int32_t fid, SLastFile *pLastF, char fname[]) {
snprintf(fname, TSDB_FILENAME_LEN - 1, "%s%s%s%sv%df%dver%" PRId64 "%s", tfsGetDiskPath(pTsdb->pVnode->pTfs, did),
TD_DIRSEP, pTsdb->path, TD_DIRSEP, TD_VID(pTsdb->pVnode), fid, pLastF->commitID, ".last");
}
void tsdbSmaFileName(STsdb *pTsdb, SDiskID did, int32_t fid, SSmaFile *pSmaF, char fname[]) {
snprintf(fname, TSDB_FILENAME_LEN - 1, "%s%s%s%sv%df%dver%" PRId64 "%s", tfsGetDiskPath(pTsdb->pVnode->pTfs, did),
TD_DIRSEP, pTsdb->path, TD_DIRSEP, TD_VID(pTsdb->pVnode), fid, pSmaF->commitID, ".sma");
}
bool tsdbDelFileIsSame(SDelFile *pDelFile1, SDelFile *pDelFile2) { return pDelFile1->commitID == pDelFile2->commitID; }
int32_t tsdbUpdateDFileHdr(TdFilePtr pFD, SDFileSet *pSet, EDataFileT ftype) {
int32_t code = 0;
int64_t n;
char hdr[TSDB_FHDR_SIZE];
memset(hdr, 0, TSDB_FHDR_SIZE);
tPutDataFileHdr(hdr, pSet, ftype);
taosCalcChecksumAppend(0, hdr, TSDB_FHDR_SIZE);
n = taosLSeekFile(pFD, 0, SEEK_SET);
if (n < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _exit;
}
n = taosWriteFile(pFD, hdr, TSDB_FHDR_SIZE);
if (n < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _exit;
}
_exit:
return code;
}
int32_t tsdbDFileRollback(STsdb *pTsdb, SDFileSet *pSet, EDataFileT ftype) {
int32_t code = 0;
int64_t size;
int64_t n;
TdFilePtr pFD;
char fname[TSDB_FILENAME_LEN];
char hdr[TSDB_FHDR_SIZE] = {0};
tsdbDataFileName(pTsdb, pSet, ftype, fname);
// truncate
switch (ftype) {
case TSDB_DATA_FILE:
size = pSet->pDataF->size;
tsdbDataFileName(pTsdb, pSet->diskId, pSet->fid, pSet->pDataF, fname);
tPutDataFile(hdr, pSet->pDataF);
break;
case TSDB_SMA_FILE:
size = pSet->pSmaF->size;
tsdbSmaFileName(pTsdb, pSet->diskId, pSet->fid, pSet->pSmaF, fname);
tPutSmaFile(hdr, pSet->pSmaF);
break;
default:
ASSERT(0);
}
taosCalcChecksumAppend(0, hdr, TSDB_FHDR_SIZE);
// open
pFD = taosOpenFile(fname, TD_FILE_WRITE);
@ -182,31 +145,24 @@ int32_t tsdbDFileRollback(STsdb *pTsdb, SDFileSet *pSet, EDataFileT ftype) {
goto _err;
}
// truncate
switch (ftype) {
case TSDB_HEAD_FILE:
size = pSet->fHead.size;
break;
case TSDB_DATA_FILE:
size = pSet->fData.size;
break;
case TSDB_LAST_FILE:
size = pSet->fLast.size;
break;
case TSDB_SMA_FILE:
size = pSet->fSma.size;
break;
default:
ASSERT(0);
}
// ftruncate
if (taosFtruncateFile(pFD, size) < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
}
// update header
code = tsdbUpdateDFileHdr(pFD, pSet, ftype);
if (code) goto _err;
n = taosLSeekFile(pFD, 0, SEEK_SET);
if (n < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
}
n = taosWriteFile(pFD, hdr, TSDB_FHDR_SIZE);
if (n < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
}
// sync
if (taosFsyncFile(pFD) < 0) {
@ -220,42 +176,20 @@ int32_t tsdbDFileRollback(STsdb *pTsdb, SDFileSet *pSet, EDataFileT ftype) {
return code;
_err:
tsdbError("vgId:%d tsdb rollback file failed since %s", TD_VID(pTsdb->pVnode), tstrerror(code));
return code;
}
int32_t tPutDataFileHdr(uint8_t *p, SDFileSet *pSet, EDataFileT ftype) {
int32_t n = 0;
switch (ftype) {
case TSDB_HEAD_FILE:
n += tPutHeadFile(p ? p + n : p, &pSet->fHead);
break;
case TSDB_DATA_FILE:
n += tPutDataFile(p ? p + n : p, &pSet->fData);
break;
case TSDB_LAST_FILE:
n += tPutLastFile(p ? p + n : p, &pSet->fLast);
break;
case TSDB_SMA_FILE:
n += tPutSmaFile(p ? p + n : p, &pSet->fSma);
break;
default:
ASSERT(0);
}
return n;
}
int32_t tPutDFileSet(uint8_t *p, SDFileSet *pSet) {
int32_t n = 0;
n += tPutI32v(p ? p + n : p, pSet->diskId.level);
n += tPutI32v(p ? p + n : p, pSet->diskId.id);
n += tPutI32v(p ? p + n : p, pSet->fid);
n += tPutHeadFile(p ? p + n : p, &pSet->fHead);
n += tPutDataFile(p ? p + n : p, &pSet->fData);
n += tPutLastFile(p ? p + n : p, &pSet->fLast);
n += tPutSmaFile(p ? p + n : p, &pSet->fSma);
n += tPutHeadFile(p ? p + n : p, pSet->pHeadF);
n += tPutDataFile(p ? p + n : p, pSet->pDataF);
n += tPutLastFile(p ? p + n : p, pSet->pLastF);
n += tPutSmaFile(p ? p + n : p, pSet->pSmaF);
return n;
}
@ -266,20 +200,18 @@ int32_t tGetDFileSet(uint8_t *p, SDFileSet *pSet) {
n += tGetI32v(p + n, &pSet->diskId.level);
n += tGetI32v(p + n, &pSet->diskId.id);
n += tGetI32v(p + n, &pSet->fid);
n += tGetHeadFile(p + n, &pSet->fHead);
n += tGetDataFile(p + n, &pSet->fData);
n += tGetLastFile(p + n, &pSet->fLast);
n += tGetSmaFile(p + n, &pSet->fSma);
n += tGetHeadFile(p + n, pSet->pHeadF);
n += tGetDataFile(p + n, pSet->pDataF);
n += tGetLastFile(p + n, pSet->pLastF);
n += tGetSmaFile(p + n, pSet->pSmaF);
return n;
}
// SDelFile ===============================================
void tsdbDelFileName(STsdb *pTsdb, SDelFile *pFile, char fname[]) {
STfs *pTfs = pTsdb->pVnode->pTfs;
snprintf(fname, TSDB_FILENAME_LEN - 1, "%s%s%s%sv%dver%" PRId64 "%s", tfsGetPrimaryPath(pTfs), TD_DIRSEP, pTsdb->path,
TD_DIRSEP, TD_VID(pTsdb->pVnode), pFile->commitID, ".del");
snprintf(fname, TSDB_FILENAME_LEN - 1, "%s%s%s%sv%dver%" PRId64 "%s", tfsGetPrimaryPath(pTsdb->pVnode->pTfs),
TD_DIRSEP, pTsdb->path, TD_DIRSEP, TD_VID(pTsdb->pVnode), pFile->commitID, ".del");
}
int32_t tPutDelFile(uint8_t *p, SDelFile *pDelFile) {

58
source/dnode/vnode/src/tsdb/tsdbMemTable.c
View File

@ -93,7 +93,11 @@ static int32_t tbDataPCmprFn(const void *p1, const void *p2) {
}
void tsdbGetTbDataFromMemTable(SMemTable *pMemTable, tb_uid_t suid, tb_uid_t uid, STbData **ppTbData) {
STbData *pTbData = &(STbData){.suid = suid, .uid = uid};
void *p = taosArraySearch(pMemTable->aTbData, &pTbData, tbDataPCmprFn, TD_EQ);
taosRLockLatch(&pMemTable->latch);
void *p = taosArraySearch(pMemTable->aTbData, &pTbData, tbDataPCmprFn, TD_EQ);
taosRUnLockLatch(&pMemTable->latch);
*ppTbData = p ? *(STbData **)p : NULL;
}
@ -363,10 +367,13 @@ static int32_t tsdbGetOrCreateTbData(SMemTable *pMemTable, tb_uid_t suid, tb_uid
void *p;
if (idx < 0) {
p = taosArrayPush(pMemTable->aTbData, &pTbData);
} else {
p = taosArrayInsert(pMemTable->aTbData, idx, &pTbData);
idx = taosArrayGetSize(pMemTable->aTbData);
}
taosWLockLatch(&pMemTable->latch);
p = taosArrayInsert(pMemTable->aTbData, idx, &pTbData);
taosWUnLockLatch(&pMemTable->latch);
if (p == NULL) {
code = TSDB_CODE_OUT_OF_MEMORY;
goto _err;
@ -605,46 +612,3 @@ void tsdbUnrefMemTable(SMemTable *pMemTable) {
tsdbMemTableDestroy(pMemTable);
}
}
int32_t tsdbTakeMemSnapshot(STsdb *pTsdb, SMemTable **ppMem, SMemTable **ppIMem) {
int32_t code = 0;
// lock
code = taosThreadRwlockRdlock(&pTsdb->rwLock);
if (code) {
code = TAOS_SYSTEM_ERROR(code);
goto _exit;
}
// take snapshot
*ppMem = pTsdb->mem;
*ppIMem = pTsdb->imem;
if (*ppMem) {
tsdbRefMemTable(*ppMem);
}
if (*ppIMem) {
tsdbRefMemTable(*ppIMem);
}
// unlock
code = taosThreadRwlockUnlock(&pTsdb->rwLock);
if (code) {
code = TAOS_SYSTEM_ERROR(code);
goto _exit;
}
_exit:
return code;
}
void tsdbUntakeMemSnapshot(STsdb *pTsdb, SMemTable *pMem, SMemTable *pIMem) {
if (pMem) {
tsdbUnrefMemTable(pMem);
}
if (pIMem) {
tsdbUnrefMemTable(pIMem);
}
}

4
source/dnode/vnode/src/tsdb/tsdbOpen.c
View File

@ -66,7 +66,7 @@ int tsdbOpen(SVnode *pVnode, STsdb **ppTsdb, const char *dir, STsdbKeepCfg *pKee
tfsMkdir(pVnode->pTfs, pTsdb->path);
// open tsdb
if (tsdbFSOpen(pTsdb, &pTsdb->pFS) < 0) {
if (tsdbFSOpen(pTsdb) < 0) {
goto _err;
}
@ -88,7 +88,7 @@ _err:
int tsdbClose(STsdb **pTsdb) {
if (*pTsdb) {
taosThreadRwlockDestroy(&(*pTsdb)->rwLock);
tsdbFSClose((*pTsdb)->pFS);
tsdbFSClose(*pTsdb);
tsdbCloseCache((*pTsdb)->lruCache);
taosMemoryFreeClear(*pTsdb);
}

122
source/dnode/vnode/src/tsdb/tsdbRead.c
View File

@ -118,8 +118,7 @@ struct STsdbReader {
char* idStr; // query info handle, for debug purpose
int32_t type; // query type: 1. retrieve all data blocks, 2. retrieve direct prev|next rows
SBlockLoadSuppInfo suppInfo;
SMemTable* pMem;
SMemTable* pIMem;
STsdbReadSnap* pReadSnap;
SIOCostSummary cost;
STSchema* pSchema;
@ -275,20 +274,18 @@ static void limitOutputBufferSize(const SQueryTableDataCond* pCond, int32_t* cap
}
// init file iterator
static int32_t initFilesetIterator(SFilesetIter* pIter, const STsdbFSState* pFState, int32_t order, const char* idstr) {
size_t numOfFileset = taosArrayGetSize(pFState->aDFileSet);
static int32_t initFilesetIterator(SFilesetIter* pIter, SArray* aDFileSet, int32_t order, const char* idstr) {
size_t numOfFileset = taosArrayGetSize(aDFileSet);
pIter->index = ASCENDING_TRAVERSE(order) ? -1 : numOfFileset;
pIter->order = order;
pIter->pFileList = taosArrayDup(pFState->aDFileSet);
pIter->pFileList = aDFileSet;
pIter->numOfFiles = numOfFileset;
tsdbDebug("init fileset iterator, total files:%d %s", pIter->numOfFiles, idstr);
return TSDB_CODE_SUCCESS;
}
static void cleanupFilesetIterator(SFilesetIter* pIter) { taosArrayDestroy(pIter->pFileList); }
static bool filesetIteratorNext(SFilesetIter* pIter, STsdbReader* pReader) {
bool asc = ASCENDING_TRAVERSE(pIter->order);
int32_t step = asc ? 1 : -1;
@ -788,11 +785,11 @@ static int32_t copyBlockDataToSDataBlock(STsdbReader* pReader, STableBlockScanIn
doCopyColVal(pColData, rowIndex++, i, &cv, pSupInfo);
}
colIndex += 1;
ASSERT(rowIndex == remain);
} else { // the specified column does not exist in file block, fill with null data
colDataAppendNNULL(pColData, 0, remain);
}
ASSERT(rowIndex == remain);
i += 1;
}
@ -1881,8 +1878,8 @@ static int32_t initMemDataIterator(STableBlockScanInfo* pBlockScanInfo, STsdbRea
int32_t backward = (!ASCENDING_TRAVERSE(pReader->order));
STbData* d = NULL;
if (pReader->pMem != NULL) {
tsdbGetTbDataFromMemTable(pReader->pMem, pReader->suid, pBlockScanInfo->uid, &d);
if (pReader->pReadSnap->pMem != NULL) {
tsdbGetTbDataFromMemTable(pReader->pReadSnap->pMem, pReader->suid, pBlockScanInfo->uid, &d);
if (d != NULL) {
code = tsdbTbDataIterCreate(d, &startKey, backward, &pBlockScanInfo->iter.iter);
if (code == TSDB_CODE_SUCCESS) {
@ -1902,8 +1899,8 @@ static int32_t initMemDataIterator(STableBlockScanInfo* pBlockScanInfo, STsdbRea
}
STbData* di = NULL;
if (pReader->pIMem != NULL) {
tsdbGetTbDataFromMemTable(pReader->pIMem, pReader->suid, pBlockScanInfo->uid, &di);
if (pReader->pReadSnap->pIMem != NULL) {
tsdbGetTbDataFromMemTable(pReader->pReadSnap->pIMem, pReader->suid, pBlockScanInfo->uid, &di);
if (di != NULL) {
code = tsdbTbDataIterCreate(di, &startKey, backward, &pBlockScanInfo->iiter.iter);
if (code == TSDB_CODE_SUCCESS) {
@ -1939,21 +1936,24 @@ int32_t initDelSkylineIterator(STableBlockScanInfo* pBlockScanInfo, STsdbReader*
SArray* pDelData = taosArrayInit(4, sizeof(SDelData));
SDelFile* pDelFile = tsdbFSStateGetDelFile(pTsdb->pFS->cState);
SDelFile* pDelFile = pReader->pReadSnap->fs.pDelFile;
if (pDelFile) {
SDelFReader* pDelFReader = NULL;
code = tsdbDelFReaderOpen(&pDelFReader, pDelFile, pTsdb, NULL);
if (code) {
if (code != TSDB_CODE_SUCCESS) {
goto _err;
}
SArray* aDelIdx = taosArrayInit(4, sizeof(SDelIdx));
if (aDelIdx == NULL) {
tsdbDelFReaderClose(&pDelFReader);
goto _err;
}
code = tsdbReadDelIdx(pDelFReader, aDelIdx, NULL);
if (code) {
if (code != TSDB_CODE_SUCCESS) {
taosArrayDestroy(aDelIdx);
tsdbDelFReaderClose(&pDelFReader);
goto _err;
}
@ -1962,9 +1962,13 @@ int32_t initDelSkylineIterator(STableBlockScanInfo* pBlockScanInfo, STsdbReader*
if (pIdx != NULL) {
code = tsdbReadDelData(pDelFReader, pIdx, pDelData, NULL);
if (code != TSDB_CODE_SUCCESS) {
goto _err;
}
}
taosArrayDestroy(aDelIdx);
tsdbDelFReaderClose(&pDelFReader);
if (code != TSDB_CODE_SUCCESS) {
goto _err;
}
}
@ -2532,8 +2536,7 @@ static int32_t checkForNeighborFileBlock(STsdbReader* pReader, STableBlockScanIn
pDumpInfo->rowIndex =
doMergeRowsInFileBlockImpl(pBlockData, pDumpInfo->rowIndex, key, pMerger, &pReader->verRange, step);
if (pDumpInfo->rowIndex >= pBlock->nRow) {
if (pDumpInfo->rowIndex >= pDumpInfo->totalRows) {
*state = CHECK_FILEBLOCK_CONT;
}
}
@ -2830,8 +2833,10 @@ int32_t tsdbReaderOpen(SVnode* pVnode, SQueryTableDataCond* pCond, SArray* pTabl
SDataBlockIter* pBlockIter = &pReader->status.blockIter;
STsdbFSState* pFState = pReader->pTsdb->pFS->cState;
initFilesetIterator(&pReader->status.fileIter, pFState, pReader->order, pReader->idStr);
code = tsdbTakeReadSnap(pReader->pTsdb, &pReader->pReadSnap);
if (code) goto _err;
initFilesetIterator(&pReader->status.fileIter, (*ppReader)->pReadSnap->fs.aDFileSet, pReader->order, pReader->idStr);
resetDataBlockIterator(&pReader->status.blockIter, pReader->order);
// no data in files, let's try buffer in memory
@ -2844,8 +2849,6 @@ int32_t tsdbReaderOpen(SVnode* pVnode, SQueryTableDataCond* pCond, SArray* pTabl
}
}
tsdbTakeMemSnapshot(pReader->pTsdb, &pReader->pMem, &pReader->pIMem);
tsdbDebug("%p total numOfTable:%d in this query %s", pReader, numOfTables, pReader->idStr);
return code;
@ -2861,7 +2864,7 @@ void tsdbReaderClose(STsdbReader* pReader) {
SBlockLoadSuppInfo* pSupInfo = &pReader->suppInfo;
tsdbUntakeMemSnapshot(pReader->pTsdb, pReader->pMem, pReader->pIMem);
tsdbUntakeReadSnap(pReader->pTsdb, pReader->pReadSnap);
taosMemoryFreeClear(pSupInfo->plist);
taosMemoryFree(pSupInfo->colIds);
@ -2874,7 +2877,6 @@ void tsdbReaderClose(STsdbReader* pReader) {
}
taosMemoryFree(pSupInfo->buildBuf);
cleanupFilesetIterator(&pReader->status.fileIter);
cleanupDataBlockIterator(&pReader->status.blockIter);
destroyBlockScanInfo(pReader->status.pTableMap);
blockDataDestroy(pReader->pResBlock);
@ -3081,8 +3083,7 @@ int32_t tsdbReaderReset(STsdbReader* pReader, SQueryTableDataCond* pCond) {
tsdbDataFReaderClose(&pReader->pFileReader);
STsdbFSState* pFState = pReader->pTsdb->pFS->cState;
initFilesetIterator(&pReader->status.fileIter, pFState, pReader->order, pReader->idStr);
initFilesetIterator(&pReader->status.fileIter, pReader->pReadSnap->fs.aDFileSet, pReader->order, pReader->idStr);
resetDataBlockIterator(&pReader->status.blockIter, pReader->order);
resetDataBlockScanInfo(pReader->status.pTableMap);
@ -3244,3 +3245,68 @@ int32_t tsdbGetTableSchema(SVnode* pVnode, int64_t uid, STSchema** pSchema, int6
return TSDB_CODE_SUCCESS;
}
int32_t tsdbTakeReadSnap(STsdb* pTsdb, STsdbReadSnap** ppSnap) {
int32_t code = 0;
// alloc
*ppSnap = (STsdbReadSnap*)taosMemoryCalloc(1, sizeof(STsdbReadSnap));
if (*ppSnap == NULL) {
code = TSDB_CODE_OUT_OF_MEMORY;
goto _exit;
}
// lock
code = taosThreadRwlockRdlock(&pTsdb->rwLock);
if (code) {
code = TAOS_SYSTEM_ERROR(code);
goto _exit;
}
// take snapshot
(*ppSnap)->pMem = pTsdb->mem;
(*ppSnap)->pIMem = pTsdb->imem;
if ((*ppSnap)->pMem) {
tsdbRefMemTable((*ppSnap)->pMem);
}
if ((*ppSnap)->pIMem) {
tsdbRefMemTable((*ppSnap)->pIMem);
}
// fs
code = tsdbFSRef(pTsdb, &(*ppSnap)->fs);
if (code) {
taosThreadRwlockUnlock(&pTsdb->rwLock);
goto _exit;
}
// unlock
code = taosThreadRwlockUnlock(&pTsdb->rwLock);
if (code) {
code = TAOS_SYSTEM_ERROR(code);
goto _exit;
}
tsdbTrace("vgId:%d take read snapshot", TD_VID(pTsdb->pVnode));
_exit:
return code;
}
void tsdbUntakeReadSnap(STsdb* pTsdb, STsdbReadSnap* pSnap) {
if (pSnap) {
if (pSnap->pMem) {
tsdbUnrefMemTable(pSnap->pMem);
}
if (pSnap->pIMem) {
tsdbUnrefMemTable(pSnap->pIMem);
}
tsdbFSUnref(pTsdb, &pSnap->fs);
taosMemoryFree(pSnap);
}
tsdbTrace("vgId:%d untake read snapshot", TD_VID(pTsdb->pVnode));
}

180
source/dnode/vnode/src/tsdb/tsdbReaderWriter.c
View File

@ -459,7 +459,7 @@ int32_t tsdbDataFReaderOpen(SDataFReader **ppReader, STsdb *pTsdb, SDFileSet *pS
// open impl
// head
tsdbDataFileName(pTsdb, pSet, TSDB_HEAD_FILE, fname);
tsdbHeadFileName(pTsdb, pSet->diskId, pSet->fid, pSet->pHeadF, fname);
pReader->pHeadFD = taosOpenFile(fname, TD_FILE_READ);
if (pReader->pHeadFD == NULL) {
code = TAOS_SYSTEM_ERROR(errno);
@ -467,7 +467,7 @@ int32_t tsdbDataFReaderOpen(SDataFReader **ppReader, STsdb *pTsdb, SDFileSet *pS
}
// data
tsdbDataFileName(pTsdb, pSet, TSDB_DATA_FILE, fname);
tsdbDataFileName(pTsdb, pSet->diskId, pSet->fid, pSet->pDataF, fname);
pReader->pDataFD = taosOpenFile(fname, TD_FILE_READ);
if (pReader->pDataFD == NULL) {
code = TAOS_SYSTEM_ERROR(errno);
@ -475,7 +475,7 @@ int32_t tsdbDataFReaderOpen(SDataFReader **ppReader, STsdb *pTsdb, SDFileSet *pS
}
// last
tsdbDataFileName(pTsdb, pSet, TSDB_LAST_FILE, fname);
tsdbLastFileName(pTsdb, pSet->diskId, pSet->fid, pSet->pLastF, fname);
pReader->pLastFD = taosOpenFile(fname, TD_FILE_READ);
if (pReader->pLastFD == NULL) {
code = TAOS_SYSTEM_ERROR(errno);
@ -483,7 +483,7 @@ int32_t tsdbDataFReaderOpen(SDataFReader **ppReader, STsdb *pTsdb, SDFileSet *pS
}
// sma
tsdbDataFileName(pTsdb, pSet, TSDB_SMA_FILE, fname);
tsdbSmaFileName(pTsdb, pSet->diskId, pSet->fid, pSet->pSmaF, fname);
pReader->pSmaFD = taosOpenFile(fname, TD_FILE_READ);
if (pReader->pSmaFD == NULL) {
code = TAOS_SYSTEM_ERROR(errno);
@ -536,8 +536,8 @@ _err:
int32_t tsdbReadBlockIdx(SDataFReader *pReader, SArray *aBlockIdx, uint8_t **ppBuf) {
int32_t code = 0;
int64_t offset = pReader->pSet->fHead.offset;
int64_t size = pReader->pSet->fHead.size - offset;
int64_t offset = pReader->pSet->pHeadF->offset;
int64_t size = pReader->pSet->pHeadF->size - offset;
uint8_t *pBuf = NULL;
int64_t n;
uint32_t delimiter;
@ -1211,17 +1211,6 @@ _err:
}
// SDataFWriter ====================================================
struct SDataFWriter {
STsdb *pTsdb;
SDFileSet wSet;
TdFilePtr pHeadFD;
TdFilePtr pDataFD;
TdFilePtr pLastFD;
TdFilePtr pSmaFD;
};
SDFileSet *tsdbDataFWriterGetWSet(SDataFWriter *pWriter) { return &pWriter->wSet; }
int32_t tsdbDataFWriterOpen(SDataFWriter **ppWriter, STsdb *pTsdb, SDFileSet *pSet) {
int32_t code = 0;
int32_t flag;
@ -1237,12 +1226,20 @@ int32_t tsdbDataFWriterOpen(SDataFWriter **ppWriter, STsdb *pTsdb, SDFileSet *pS
goto _err;
}
pWriter->pTsdb = pTsdb;
pWriter->wSet = *pSet;
pSet = &pWriter->wSet;
pWriter->wSet = (SDFileSet){.diskId = pSet->diskId,
.fid = pSet->fid,
.pHeadF = &pWriter->fHead,
.pDataF = &pWriter->fData,
.pLastF = &pWriter->fLast,
.pSmaF = &pWriter->fSma};
pWriter->fHead = *pSet->pHeadF;
pWriter->fData = *pSet->pDataF;
pWriter->fLast = *pSet->pLastF;
pWriter->fSma = *pSet->pSmaF;
// head
flag = TD_FILE_WRITE | TD_FILE_CREATE | TD_FILE_TRUNC;
tsdbDataFileName(pTsdb, pSet, TSDB_HEAD_FILE, fname);
tsdbHeadFileName(pTsdb, pWriter->wSet.diskId, pWriter->wSet.fid, &pWriter->fHead, fname);
pWriter->pHeadFD = taosOpenFile(fname, flag);
if (pWriter->pHeadFD == NULL) {
code = TAOS_SYSTEM_ERROR(errno);
@ -1257,28 +1254,28 @@ int32_t tsdbDataFWriterOpen(SDataFWriter **ppWriter, STsdb *pTsdb, SDFileSet *pS
ASSERT(n == TSDB_FHDR_SIZE);
pSet->fHead.size += TSDB_FHDR_SIZE;
pWriter->fHead.size += TSDB_FHDR_SIZE;
// data
if (pSet->fData.size == 0) {
if (pWriter->fData.size == 0) {
flag = TD_FILE_WRITE | TD_FILE_CREATE | TD_FILE_TRUNC;
} else {
flag = TD_FILE_WRITE;
}
tsdbDataFileName(pTsdb, pSet, TSDB_DATA_FILE, fname);
tsdbDataFileName(pTsdb, pWriter->wSet.diskId, pWriter->wSet.fid, &pWriter->fData, fname);
pWriter->pDataFD = taosOpenFile(fname, flag);
if (pWriter->pDataFD == NULL) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
}
if (pSet->fData.size == 0) {
if (pWriter->fData.size == 0) {
n = taosWriteFile(pWriter->pDataFD, hdr, TSDB_FHDR_SIZE);
if (n < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
}
pSet->fData.size += TSDB_FHDR_SIZE;
pWriter->fData.size += TSDB_FHDR_SIZE;
} else {
n = taosLSeekFile(pWriter->pDataFD, 0, SEEK_END);
if (n < 0) {
@ -1286,29 +1283,29 @@ int32_t tsdbDataFWriterOpen(SDataFWriter **ppWriter, STsdb *pTsdb, SDFileSet *pS
goto _err;
}
ASSERT(n == pSet->fData.size);
ASSERT(n == pWriter->fData.size);
}
// last
if (pSet->fLast.size == 0) {
if (pWriter->fLast.size == 0) {
flag = TD_FILE_WRITE | TD_FILE_CREATE | TD_FILE_TRUNC;
} else {
flag = TD_FILE_WRITE;
}
tsdbDataFileName(pTsdb, pSet, TSDB_LAST_FILE, fname);
tsdbLastFileName(pTsdb, pWriter->wSet.diskId, pWriter->wSet.fid, &pWriter->fLast, fname);
pWriter->pLastFD = taosOpenFile(fname, flag);
if (pWriter->pLastFD == NULL) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
}
if (pSet->fLast.size == 0) {
if (pWriter->fLast.size == 0) {
n = taosWriteFile(pWriter->pLastFD, hdr, TSDB_FHDR_SIZE);
if (n < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
}
pSet->fLast.size += TSDB_FHDR_SIZE;
pWriter->fLast.size += TSDB_FHDR_SIZE;
} else {
n = taosLSeekFile(pWriter->pLastFD, 0, SEEK_END);
if (n < 0) {
@ -1316,29 +1313,29 @@ int32_t tsdbDataFWriterOpen(SDataFWriter **ppWriter, STsdb *pTsdb, SDFileSet *pS
goto _err;
}
ASSERT(n == pSet->fLast.size);
ASSERT(n == pWriter->fLast.size);
}
// sma
if (pSet->fSma.size == 0) {
if (pWriter->fSma.size == 0) {
flag = TD_FILE_WRITE | TD_FILE_CREATE | TD_FILE_TRUNC;
} else {
flag = TD_FILE_WRITE;
}
tsdbDataFileName(pTsdb, pSet, TSDB_SMA_FILE, fname);
tsdbSmaFileName(pTsdb, pWriter->wSet.diskId, pWriter->wSet.fid, &pWriter->fSma, fname);
pWriter->pSmaFD = taosOpenFile(fname, flag);
if (pWriter->pSmaFD == NULL) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
}
if (pSet->fSma.size == 0) {
if (pWriter->fSma.size == 0) {
n = taosWriteFile(pWriter->pSmaFD, hdr, TSDB_FHDR_SIZE);
if (n < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
}
pSet->fSma.size += TSDB_FHDR_SIZE;
pWriter->fSma.size += TSDB_FHDR_SIZE;
} else {
n = taosLSeekFile(pWriter->pSmaFD, 0, SEEK_END);
if (n < 0) {
@ -1346,7 +1343,7 @@ int32_t tsdbDataFWriterOpen(SDataFWriter **ppWriter, STsdb *pTsdb, SDFileSet *pS
goto _err;
}
ASSERT(n == pSet->fSma.size);
ASSERT(n == pWriter->fSma.size);
}
*ppWriter = pWriter;
@ -1418,22 +1415,76 @@ _err:
int32_t tsdbUpdateDFileSetHeader(SDataFWriter *pWriter) {
int32_t code = 0;
int64_t n;
char hdr[TSDB_FHDR_SIZE];
// head ==============
code = tsdbUpdateDFileHdr(pWriter->pHeadFD, &pWriter->wSet, TSDB_HEAD_FILE);
if (code) goto _err;
memset(hdr, 0, TSDB_FHDR_SIZE);
tPutHeadFile(hdr, &pWriter->fHead);
taosCalcChecksumAppend(0, hdr, TSDB_FHDR_SIZE);
n = taosLSeekFile(pWriter->pHeadFD, 0, SEEK_SET);
if (n < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
}
n = taosWriteFile(pWriter->pHeadFD, hdr, TSDB_FHDR_SIZE);
if (n < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
}
// data ==============
code = tsdbUpdateDFileHdr(pWriter->pHeadFD, &pWriter->wSet, TSDB_DATA_FILE);
if (code) goto _err;
memset(hdr, 0, TSDB_FHDR_SIZE);
tPutDataFile(hdr, &pWriter->fData);
taosCalcChecksumAppend(0, hdr, TSDB_FHDR_SIZE);
n = taosLSeekFile(pWriter->pDataFD, 0, SEEK_SET);
if (n < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
}
n = taosWriteFile(pWriter->pDataFD, hdr, TSDB_FHDR_SIZE);
if (n < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
}
// last ==============
code = tsdbUpdateDFileHdr(pWriter->pHeadFD, &pWriter->wSet, TSDB_LAST_FILE);
if (code) goto _err;
memset(hdr, 0, TSDB_FHDR_SIZE);
tPutLastFile(hdr, &pWriter->fLast);
taosCalcChecksumAppend(0, hdr, TSDB_FHDR_SIZE);
n = taosLSeekFile(pWriter->pLastFD, 0, SEEK_SET);
if (n < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
}
n = taosWriteFile(pWriter->pLastFD, hdr, TSDB_FHDR_SIZE);
if (n < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
}
// sma ==============
code = tsdbUpdateDFileHdr(pWriter->pHeadFD, &pWriter->wSet, TSDB_SMA_FILE);
if (code) goto _err;
memset(hdr, 0, TSDB_FHDR_SIZE);
tPutSmaFile(hdr, &pWriter->fSma);
taosCalcChecksumAppend(0, hdr, TSDB_FHDR_SIZE);
n = taosLSeekFile(pWriter->pSmaFD, 0, SEEK_SET);
if (n < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
}
n = taosWriteFile(pWriter->pSmaFD, hdr, TSDB_FHDR_SIZE);
if (n < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
}
return code;
@ -1444,7 +1495,7 @@ _err:
int32_t tsdbWriteBlockIdx(SDataFWriter *pWriter, SArray *aBlockIdx, uint8_t **ppBuf) {
int32_t code = 0;
SHeadFile *pHeadFile = &pWriter->wSet.fHead;
SHeadFile *pHeadFile = &pWriter->fHead;
uint8_t *pBuf = NULL;
int64_t size;
int64_t n;
@ -1494,7 +1545,7 @@ _err:
int32_t tsdbWriteBlock(SDataFWriter *pWriter, SMapData *mBlock, uint8_t **ppBuf, SBlockIdx *pBlockIdx) {
int32_t code = 0;
SHeadFile *pHeadFile = &pWriter->wSet.fHead;
SHeadFile *pHeadFile = &pWriter->fHead;
SBlockDataHdr hdr = {.delimiter = TSDB_FILE_DLMT, .suid = pBlockIdx->suid, .uid = pBlockIdx->uid};
uint8_t *pBuf = NULL;
int64_t size;
@ -1831,9 +1882,9 @@ int32_t tsdbWriteBlockData(SDataFWriter *pWriter, SBlockData *pBlockData, uint8_
pSubBlock->nRow = pBlockData->nRow;
pSubBlock->cmprAlg = cmprAlg;
if (pBlock->last) {
pSubBlock->offset = pWriter->wSet.fLast.size;
pSubBlock->offset = pWriter->fLast.size;
} else {
pSubBlock->offset = pWriter->wSet.fData.size;
pSubBlock->offset = pWriter->fData.size;
}
// ======================= BLOCK DATA =======================
@ -1881,9 +1932,9 @@ int32_t tsdbWriteBlockData(SDataFWriter *pWriter, SBlockData *pBlockData, uint8_
pSubBlock->szBlock = pSubBlock->szBlockCol + sizeof(TSCKSUM) + nData;
if (pBlock->last) {
pWriter->wSet.fLast.size += pSubBlock->szBlock;
pWriter->fLast.size += pSubBlock->szBlock;
} else {
pWriter->wSet.fData.size += pSubBlock->szBlock;
pWriter->fData.size += pSubBlock->szBlock;
}
// ======================= BLOCK SMA =======================
@ -1896,8 +1947,8 @@ int32_t tsdbWriteBlockData(SDataFWriter *pWriter, SBlockData *pBlockData, uint8_
if (code) goto _err;
if (pSubBlock->nSma > 0) {
pSubBlock->sOffset = pWriter->wSet.fSma.size;
pWriter->wSet.fSma.size += (sizeof(SColumnDataAgg) * pSubBlock->nSma + sizeof(TSCKSUM));
pSubBlock->sOffset = pWriter->fSma.size;
pWriter->fSma.size += (sizeof(SColumnDataAgg) * pSubBlock->nSma + sizeof(TSCKSUM));
}
_exit:
@ -1924,8 +1975,8 @@ int32_t tsdbDFileSetCopy(STsdb *pTsdb, SDFileSet *pSetFrom, SDFileSet *pSetTo) {
char fNameTo[TSDB_FILENAME_LEN];
// head
tsdbDataFileName(pTsdb, pSetFrom, TSDB_HEAD_FILE, fNameFrom);
tsdbDataFileName(pTsdb, pSetTo, TSDB_HEAD_FILE, fNameTo);
tsdbHeadFileName(pTsdb, pSetFrom->diskId, pSetFrom->fid, pSetFrom->pHeadF, fNameFrom);
tsdbHeadFileName(pTsdb, pSetTo->diskId, pSetTo->fid, pSetTo->pHeadF, fNameTo);
pOutFD = taosOpenFile(fNameTo, TD_FILE_WRITE | TD_FILE_CREATE | TD_FILE_TRUNC);
if (pOutFD == NULL) {
@ -1939,7 +1990,7 @@ int32_t tsdbDFileSetCopy(STsdb *pTsdb, SDFileSet *pSetFrom, SDFileSet *pSetTo) {
goto _err;
}
n = taosFSendFile(pOutFD, PInFD, 0, pSetFrom->fHead.size);
n = taosFSendFile(pOutFD, PInFD, 0, pSetFrom->pHeadF->size);
if (n < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
@ -1948,8 +1999,8 @@ int32_t tsdbDFileSetCopy(STsdb *pTsdb, SDFileSet *pSetFrom, SDFileSet *pSetTo) {
taosCloseFile(&PInFD);
// data
tsdbDataFileName(pTsdb, pSetFrom, TSDB_DATA_FILE, fNameFrom);
tsdbDataFileName(pTsdb, pSetTo, TSDB_DATA_FILE, fNameTo);
tsdbDataFileName(pTsdb, pSetFrom->diskId, pSetFrom->fid, pSetFrom->pDataF, fNameFrom);
tsdbDataFileName(pTsdb, pSetTo->diskId, pSetTo->fid, pSetTo->pDataF, fNameTo);
pOutFD = taosOpenFile(fNameTo, TD_FILE_WRITE | TD_FILE_CREATE | TD_FILE_TRUNC);
if (pOutFD == NULL) {
@ -1963,7 +2014,7 @@ int32_t tsdbDFileSetCopy(STsdb *pTsdb, SDFileSet *pSetFrom, SDFileSet *pSetTo) {
goto _err;
}
n = taosFSendFile(pOutFD, PInFD, 0, pSetFrom->fData.size);
n = taosFSendFile(pOutFD, PInFD, 0, pSetFrom->pDataF->size);
if (n < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
@ -1972,8 +2023,9 @@ int32_t tsdbDFileSetCopy(STsdb *pTsdb, SDFileSet *pSetFrom, SDFileSet *pSetTo) {
taosCloseFile(&PInFD);
// last
tsdbDataFileName(pTsdb, pSetFrom, TSDB_LAST_FILE, fNameFrom);
tsdbDataFileName(pTsdb, pSetTo, TSDB_LAST_FILE, fNameTo);
tsdbLastFileName(pTsdb, pSetFrom->diskId, pSetFrom->fid, pSetFrom->pLastF, fNameFrom);
tsdbLastFileName(pTsdb, pSetTo->diskId, pSetTo->fid, pSetTo->pLastF, fNameTo);
pOutFD = taosOpenFile(fNameTo, TD_FILE_WRITE | TD_FILE_CREATE | TD_FILE_TRUNC);
if (pOutFD == NULL) {
code = TAOS_SYSTEM_ERROR(errno);
@ -1986,7 +2038,7 @@ int32_t tsdbDFileSetCopy(STsdb *pTsdb, SDFileSet *pSetFrom, SDFileSet *pSetTo) {
goto _err;
}
n = taosFSendFile(pOutFD, PInFD, 0, pSetFrom->fLast.size);
n = taosFSendFile(pOutFD, PInFD, 0, pSetFrom->pLastF->size);
if (n < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;
@ -1995,8 +2047,8 @@ int32_t tsdbDFileSetCopy(STsdb *pTsdb, SDFileSet *pSetFrom, SDFileSet *pSetTo) {
taosCloseFile(&PInFD);
// sma
tsdbDataFileName(pTsdb, pSetFrom, TSDB_SMA_FILE, fNameFrom);
tsdbDataFileName(pTsdb, pSetTo, TSDB_SMA_FILE, fNameTo);
tsdbSmaFileName(pTsdb, pSetFrom->diskId, pSetFrom->fid, pSetFrom->pSmaF, fNameFrom);
tsdbSmaFileName(pTsdb, pSetTo->diskId, pSetTo->fid, pSetTo->pSmaF, fNameTo);
pOutFD = taosOpenFile(fNameTo, TD_FILE_WRITE | TD_FILE_CREATE | TD_FILE_TRUNC);
if (pOutFD == NULL) {
@ -2010,7 +2062,7 @@ int32_t tsdbDFileSetCopy(STsdb *pTsdb, SDFileSet *pSetFrom, SDFileSet *pSetTo) {
goto _err;
}
n = taosFSendFile(pOutFD, PInFD, 0, pSetFrom->fSma.size);
n = taosFSendFile(pOutFD, PInFD, 0, pSetFrom->pSmaF->size);
if (n < 0) {
code = TAOS_SYSTEM_ERROR(errno);
goto _err;

129
source/dnode/vnode/src/tsdb/tsdbRetention.c
View File

@ -15,90 +15,99 @@
#include "tsdb.h"
static int32_t tsdbDoRetentionImpl(STsdb *pTsdb, int64_t now, int8_t try, int8_t *canDo) {
int32_t code = 0;
STsdbFSState *pState;
if (try) {
pState = pTsdb->pFS->cState;
*canDo = 0;
} else {
pState = pTsdb->pFS->nState;
}
for (int32_t iSet = 0; iSet < taosArrayGetSize(pState->aDFileSet); iSet++) {
SDFileSet *pDFileSet = (SDFileSet *)taosArrayGet(pState->aDFileSet, iSet);
int32_t expLevel = tsdbFidLevel(pDFileSet->fid, &pTsdb->keepCfg, now);
static bool tsdbShouldDoRetention(STsdb *pTsdb, int64_t now) {
for (int32_t iSet = 0; iSet < taosArrayGetSize(pTsdb->fs.aDFileSet); iSet++) {
SDFileSet *pSet = (SDFileSet *)taosArrayGet(pTsdb->fs.aDFileSet, iSet);
int32_t expLevel = tsdbFidLevel(pSet->fid, &pTsdb->keepCfg, now);
SDiskID did;
// check
if (expLevel == pDFileSet->diskId.id) continue;
if (expLevel == pSet->diskId.level) continue;
// delete or move
if (expLevel < 0) {
if (try) {
*canDo = 1;
} else {
tsdbFSStateDeleteDFileSet(pState, pDFileSet->fid);
iSet--;
}
return true;
} else {
if (tfsAllocDisk(pTsdb->pVnode->pTfs, expLevel, &did) < 0) {
return false;
}
if (did.level == pSet->diskId.level) continue;
return true;
}
}
return false;
}
int32_t tsdbDoRetention(STsdb *pTsdb, int64_t now) {
int32_t code = 0;
if (!tsdbShouldDoRetention(pTsdb, now)) {
return code;
}
// do retention
STsdbFS fs;
code = tsdbFSCopy(pTsdb, &fs);
if (code) goto _err;
for (int32_t iSet = 0; iSet < taosArrayGetSize(fs.aDFileSet); iSet++) {
SDFileSet *pSet = (SDFileSet *)taosArrayGet(pTsdb->fs.aDFileSet, iSet);
int32_t expLevel = tsdbFidLevel(pSet->fid, &pTsdb->keepCfg, now);
SDiskID did;
if (expLevel < 0) {
taosMemoryFree(pSet->pHeadF);
taosMemoryFree(pSet->pDataF);
taosMemoryFree(pSet->pLastF);
taosMemoryFree(pSet->pSmaF);
taosArrayRemove(fs.aDFileSet, iSet);
iSet--;
} else {
// alloc
if (tfsAllocDisk(pTsdb->pVnode->pTfs, expLevel, &did) < 0) {
code = terrno;
goto _exit;
}
if (did.level == pDFileSet->diskId.level) continue;
if (did.level == pSet->diskId.level) continue;
if (try) {
*canDo = 1;
} else {
// copy the file to new disk
// copy file to new disk (todo)
SDFileSet fSet = *pSet;
fSet.diskId = did;
SDFileSet nDFileSet = *pDFileSet;
nDFileSet.diskId = did;
code = tsdbDFileSetCopy(pTsdb, pSet, &fSet);
if (code) goto _err;
tfsMkdirRecurAt(pTsdb->pVnode->pTfs, pTsdb->path, did);
code = tsdbDFileSetCopy(pTsdb, pDFileSet, &nDFileSet);
if (code) goto _exit;
code = tsdbFSStateUpsertDFileSet(pState, &nDFileSet);
if (code) goto _exit;
}
code = tsdbFSUpsertFSet(&fs, &fSet);
if (code) goto _err;
}
/* code */
}
_exit:
return code;
}
int32_t tsdbDoRetention(STsdb *pTsdb, int64_t now) {
int32_t code = 0;
int8_t canDo;
// try
tsdbDoRetentionImpl(pTsdb, now, 1, &canDo);
if (!canDo) goto _exit;
// begin
code = tsdbFSBegin(pTsdb->pFS);
// do change fs
code = tsdbFSCommit1(pTsdb, &fs);
if (code) goto _err;
// do retention
code = tsdbDoRetentionImpl(pTsdb, now, 0, NULL);
if (code) goto _err;
taosThreadRwlockWrlock(&pTsdb->rwLock);
// commit
code = tsdbFSCommit(pTsdb->pFS);
if (code) goto _err;
code = tsdbFSCommit2(pTsdb, &fs);
if (code) {
taosThreadRwlockUnlock(&pTsdb->rwLock);
goto _err;
}
taosThreadRwlockUnlock(&pTsdb->rwLock);
tsdbFSDestroy(&fs);
_exit:
return code;
_err:
tsdbError("vgId:%d tsdb do retention failed since %s", TD_VID(pTsdb->pVnode), tstrerror(code));
tsdbFSRollback(pTsdb->pFS);
ASSERT(0);
// tsdbFSRollback(pTsdb->pFS);
return code;
}

81
source/dnode/vnode/src/tsdb/tsdbSnapshot.c
View File

@ -20,6 +20,7 @@ struct STsdbSnapReader {
STsdb* pTsdb;
int64_t sver;
int64_t ever;
STsdbFS fs;
// for data file
int8_t dataDone;
int32_t fid;
@ -45,7 +46,8 @@ static int32_t tsdbSnapReadData(STsdbSnapReader* pReader, uint8_t** ppData) {
while (true) {
if (pReader->pDataFReader == NULL) {
SDFileSet* pSet = tsdbFSStateGetDFileSet(pTsdb->pFS->cState, pReader->fid, TD_GT);
SDFileSet* pSet =
taosArraySearch(pReader->fs.aDFileSet, &(SDFileSet){.fid = pReader->fid}, tDFileSetCmprFn, TD_GT);
if (pSet == NULL) goto _exit;
@ -159,7 +161,7 @@ _err:
static int32_t tsdbSnapReadDel(STsdbSnapReader* pReader, uint8_t** ppData) {
int32_t code = 0;
STsdb* pTsdb = pReader->pTsdb;
SDelFile* pDelFile = pTsdb->pFS->cState->pDelFile;
SDelFile* pDelFile = pReader->fs.pDelFile;
if (pReader->pDelFReader == NULL) {
if (pDelFile == NULL) {
@ -254,6 +256,24 @@ int32_t tsdbSnapReaderOpen(STsdb* pTsdb, int64_t sver, int64_t ever, STsdbSnapRe
pReader->sver = sver;
pReader->ever = ever;
code = taosThreadRwlockRdlock(&pTsdb->rwLock);
if (code) {
code = TAOS_SYSTEM_ERROR(code);
goto _err;
}
code = tsdbFSRef(pTsdb, &pReader->fs);
if (code) {
taosThreadRwlockUnlock(&pTsdb->rwLock);
goto _err;
}
code = taosThreadRwlockUnlock(&pTsdb->rwLock);
if (code) {
code = TAOS_SYSTEM_ERROR(code);
goto _err;
}
pReader->fid = INT32_MIN;
pReader->aBlockIdx = taosArrayInit(0, sizeof(SBlockIdx));
if (pReader->aBlockIdx == NULL) {
@ -305,6 +325,8 @@ int32_t tsdbSnapReaderClose(STsdbSnapReader** ppReader) {
taosArrayDestroy(pReader->aDelIdx);
taosArrayDestroy(pReader->aDelData);
tsdbFSUnref(pReader->pTsdb, &pReader->fs);
tsdbInfo("vgId:%d vnode snapshot tsdb reader closed", TD_VID(pReader->pTsdb->pVnode));
taosMemoryFree(pReader);
@ -358,6 +380,7 @@ struct STsdbSnapWriter {
STsdb* pTsdb;
int64_t sver;
int64_t ever;
STsdbFS fs;
// config
int32_t minutes;
@ -798,7 +821,7 @@ static int32_t tsdbSnapWriteDataEnd(STsdbSnapWriter* pWriter) {
code = tsdbWriteBlockIdx(pWriter->pDataFWriter, pWriter->aBlockIdxW, NULL);
if (code) goto _err;
code = tsdbFSStateUpsertDFileSet(pTsdb->pFS->nState, tsdbDataFWriterGetWSet(pWriter->pDataFWriter));
code = tsdbFSUpsertFSet(&pWriter->fs, &pWriter->pDataFWriter->wSet);
if (code) goto _err;
code = tsdbDataFWriterClose(&pWriter->pDataFWriter, 1);
@ -843,7 +866,7 @@ static int32_t tsdbSnapWriteData(STsdbSnapWriter* pWriter, uint8_t* pData, uint3
pWriter->fid = fid;
// read
SDFileSet* pSet = tsdbFSStateGetDFileSet(pTsdb->pFS->nState, fid, TD_EQ);
SDFileSet* pSet = taosArraySearch(pWriter->fs.aDFileSet, &(SDFileSet){.fid = fid}, tDFileSetCmprFn, TD_EQ);
if (pSet) {
code = tsdbDataFReaderOpen(&pWriter->pDataFReader, pTsdb, pSet);
if (code) goto _err;
@ -863,22 +886,26 @@ static int32_t tsdbSnapWriteData(STsdbSnapWriter* pWriter, uint8_t* pData, uint3
tBlockDataReset(&pWriter->bDataR);
// write
SDFileSet wSet;
SHeadFile fHead;
SDataFile fData;
SLastFile fLast;
SSmaFile fSma;
SDFileSet wSet = {.pHeadF = &fHead, .pDataF = &fData, .pLastF = &fLast, .pSmaF = &fSma};
if (pSet) {
wSet = (SDFileSet){.diskId = pSet->diskId,
.fid = fid,
.fHead = {.commitID = pWriter->commitID, .offset = 0, .size = 0},
.fData = pSet->fData,
.fLast = {.commitID = pWriter->commitID, .size = 0},
.fSma = pSet->fSma};
wSet.diskId = pSet->diskId;
wSet.fid = fid;
fHead = (SHeadFile){.commitID = pWriter->commitID, .offset = 0, .size = 0};
fData = *pSet->pDataF;
fLast = (SLastFile){.commitID = pWriter->commitID, .size = 0};
fSma = *pSet->pSmaF;
} else {
wSet = (SDFileSet){.diskId = (SDiskID){.level = 0, .id = 0},
.fid = fid,
.fHead = {.commitID = pWriter->commitID, .offset = 0, .size = 0},
.fData = {.commitID = pWriter->commitID, .size = 0},
.fLast = {.commitID = pWriter->commitID, .size = 0},
.fSma = {.commitID = pWriter->commitID, .size = 0}};
wSet.diskId = (SDiskID){.level = 0, .id = 0};
wSet.fid = fid;
fHead = (SHeadFile){.commitID = pWriter->commitID, .offset = 0, .size = 0};
fData = (SDataFile){.commitID = pWriter->commitID, .size = 0};
fLast = (SLastFile){.commitID = pWriter->commitID, .size = 0};
fSma = (SSmaFile){.commitID = pWriter->commitID, .size = 0};
}
code = tsdbDataFWriterOpen(&pWriter->pDataFWriter, pTsdb, &wSet);
@ -907,7 +934,7 @@ static int32_t tsdbSnapWriteDel(STsdbSnapWriter* pWriter, uint8_t* pData, uint32
STsdb* pTsdb = pWriter->pTsdb;
if (pWriter->pDelFWriter == NULL) {
SDelFile* pDelFile = tsdbFSStateGetDelFile(pTsdb->pFS->nState);
SDelFile* pDelFile = pWriter->fs.pDelFile;
// reader
if (pDelFile) {
@ -1017,7 +1044,7 @@ static int32_t tsdbSnapWriteDelEnd(STsdbSnapWriter* pWriter) {
code = tsdbUpdateDelFileHdr(pWriter->pDelFWriter);
if (code) goto _err;
code = tsdbFSStateUpsertDelFile(pTsdb->pFS->nState, &pWriter->pDelFWriter->fDel);
code = tsdbFSUpsertDelFile(&pWriter->fs, &pWriter->pDelFWriter->fDel);
if (code) goto _err;
code = tsdbDelFWriterClose(&pWriter->pDelFWriter, 1);
@ -1051,6 +1078,9 @@ int32_t tsdbSnapWriterOpen(STsdb* pTsdb, int64_t sver, int64_t ever, STsdbSnapWr
pWriter->sver = sver;
pWriter->ever = ever;
code = tsdbFSCopy(pTsdb, &pWriter->fs);
if (code) goto _err;
// config
pWriter->minutes = pTsdb->keepCfg.days;
pWriter->precision = pTsdb->keepCfg.precision;
@ -1096,9 +1126,6 @@ int32_t tsdbSnapWriterOpen(STsdb* pTsdb, int64_t sver, int64_t ever, STsdbSnapWr
goto _err;
}
code = tsdbFSBegin(pTsdb->pFS);
if (code) goto _err;
*ppWriter = pWriter;
return code;
@ -1113,8 +1140,9 @@ int32_t tsdbSnapWriterClose(STsdbSnapWriter** ppWriter, int8_t rollback) {
STsdbSnapWriter* pWriter = *ppWriter;
if (rollback) {
code = tsdbFSRollback(pWriter->pTsdb->pFS);
if (code) goto _err;
ASSERT(0);
// code = tsdbFSRollback(pWriter->pTsdb->pFS);
// if (code) goto _err;
} else {
code = tsdbSnapWriteDataEnd(pWriter);
if (code) goto _err;
@ -1122,7 +1150,10 @@ int32_t tsdbSnapWriterClose(STsdbSnapWriter** ppWriter, int8_t rollback) {
code = tsdbSnapWriteDelEnd(pWriter);
if (code) goto _err;
code = tsdbFSCommit(pWriter->pTsdb->pFS);
code = tsdbFSCommit1(pWriter->pTsdb, &pWriter->fs);
if (code) goto _err;
code = tsdbFSCommit2(pWriter->pTsdb, &pWriter->fs);
if (code) goto _err;
}

5
source/dnode/vnode/src/vnd/vnodeSvr.c
View File

@ -316,7 +316,7 @@ int32_t vnodeProcessFetchMsg(SVnode *pVnode, SRpcMsg *pMsg, SQueueInfo *pInfo) {
case TDMT_VND_TABLE_CFG:
return vnodeGetTableCfg(pVnode, pMsg);
case TDMT_VND_CONSUME:
return tqProcessPollReq(pVnode->pTq, pMsg, pInfo->workerId);
return tqProcessPollReq(pVnode->pTq, pMsg);
case TDMT_STREAM_TASK_RUN:
return tqProcessTaskRunReq(pVnode->pTq, pMsg);
case TDMT_STREAM_TASK_DISPATCH:
@ -773,6 +773,7 @@ static int32_t vnodeProcessSubmitReq(SVnode *pVnode, int64_t version, void *pReq
terrno = TSDB_CODE_SUCCESS;
pRsp->code = 0;
pSubmitReq->version = version;
#ifdef TD_DEBUG_PRINT_ROW
vnodeDebugPrintSubmitMsg(pVnode, pReq, __func__);
@ -791,7 +792,7 @@ static int32_t vnodeProcessSubmitReq(SVnode *pVnode, int64_t version, void *pReq
submitRsp.pArray = taosArrayInit(msgIter.numOfBlocks, sizeof(SSubmitBlkRsp));
newTbUids = taosArrayInit(msgIter.numOfBlocks, sizeof(int64_t));
if (!submitRsp.pArray) {
if (!submitRsp.pArray || !newTbUids) {
pRsp->code = TSDB_CODE_OUT_OF_MEMORY;
goto _exit;
}

2
source/libs/catalog/inc/catalogInt.h
View File

@ -460,8 +460,6 @@ typedef struct SCtgOperation {
#define CTG_FLAG_MAKE_STB(_isStb) (((_isStb) == 1) ? CTG_FLAG_STB : ((_isStb) == 0 ? CTG_FLAG_NOT_STB : CTG_FLAG_UNKNOWN_STB))
#define CTG_FLAG_MATCH_STB(_flag, tbType) (CTG_FLAG_IS_UNKNOWN_STB(_flag) || (CTG_FLAG_IS_STB(_flag) && (tbType) == TSDB_SUPER_TABLE) || (CTG_FLAG_IS_NOT_STB(_flag) && (tbType) != TSDB_SUPER_TABLE))
#define CTG_IS_SYS_DBNAME(_dbname) (((*(_dbname) == 'i') && (0 == strcmp(_dbname, TSDB_INFORMATION_SCHEMA_DB))) || ((*(_dbname) == 'p') && (0 == strcmp(_dbname, TSDB_PERFORMANCE_SCHEMA_DB))))
#define CTG_META_SIZE(pMeta) (sizeof(STableMeta) + ((pMeta)->tableInfo.numOfTags + (pMeta)->tableInfo.numOfColumns) * sizeof(SSchema))
#define CTG_TABLE_NOT_EXIST(code) (code == CTG_ERR_CODE_TABLE_NOT_EXIST)

6
source/libs/catalog/src/catalog.c
View File

@ -865,7 +865,7 @@ int32_t catalogChkTbMetaVersion(SCatalog* pCtg, SRequestConnInfo *pConn, SArray*
tNameFromString(&name, pTb->tbFName, T_NAME_ACCT | T_NAME_DB | T_NAME_TABLE);
if (CTG_IS_SYS_DBNAME(name.dbname)) {
if (IS_SYS_DBNAME(name.dbname)) {
continue;
}
@ -936,7 +936,7 @@ int32_t catalogGetTableDistVgInfo(SCatalog* pCtg, SRequestConnInfo *pConn, const
CTG_API_LEAVE(TSDB_CODE_CTG_INVALID_INPUT);
}
if (CTG_IS_SYS_DBNAME(pTableName->dbname)) {
if (IS_SYS_DBNAME(pTableName->dbname)) {
ctgError("no valid vgInfo for db, dbname:%s", pTableName->dbname);
CTG_API_LEAVE(TSDB_CODE_CTG_INVALID_INPUT);
}
@ -947,7 +947,7 @@ int32_t catalogGetTableDistVgInfo(SCatalog* pCtg, SRequestConnInfo *pConn, const
int32_t catalogGetTableHashVgroup(SCatalog *pCtg, SRequestConnInfo *pConn, const SName *pTableName, SVgroupInfo *pVgroup) {
CTG_API_ENTER();
if (CTG_IS_SYS_DBNAME(pTableName->dbname)) {
if (IS_SYS_DBNAME(pTableName->dbname)) {
ctgError("no valid vgInfo for db, dbname:%s", pTableName->dbname);
CTG_API_LEAVE(TSDB_CODE_CTG_INVALID_INPUT);
}

14
source/libs/catalog/src/ctgCache.c
View File

@ -132,7 +132,7 @@ void ctgReleaseDBCache(SCatalog *pCtg, SCtgDBCache *dbCache) {
int32_t ctgAcquireDBCacheImpl(SCatalog* pCtg, const char *dbFName, SCtgDBCache **pCache, bool acquire) {
char *p = strchr(dbFName, '.');
if (p && CTG_IS_SYS_DBNAME(p + 1)) {
if (p && IS_SYS_DBNAME(p + 1)) {
dbFName = p + 1;
}
@ -694,7 +694,7 @@ int32_t ctgDropDbCacheEnqueue(SCatalog* pCtg, const char *dbFName, int64_t dbId)
}
char *p = strchr(dbFName, '.');
if (p && CTG_IS_SYS_DBNAME(p + 1)) {
if (p && IS_SYS_DBNAME(p + 1)) {
dbFName = p + 1;
}
@ -727,7 +727,7 @@ int32_t ctgDropDbVgroupEnqueue(SCatalog* pCtg, const char *dbFName, bool syncOp)
}
char *p = strchr(dbFName, '.');
if (p && CTG_IS_SYS_DBNAME(p + 1)) {
if (p && IS_SYS_DBNAME(p + 1)) {
dbFName = p + 1;
}
@ -823,7 +823,7 @@ int32_t ctgUpdateVgroupEnqueue(SCatalog* pCtg, const char *dbFName, int64_t dbId
}
char *p = strchr(dbFName, '.');
if (p && CTG_IS_SYS_DBNAME(p + 1)) {
if (p && IS_SYS_DBNAME(p + 1)) {
dbFName = p + 1;
}
@ -859,7 +859,7 @@ int32_t ctgUpdateTbMetaEnqueue(SCatalog* pCtg, STableMetaOutput *output, bool sy
}
char *p = strchr(output->dbFName, '.');
if (p && CTG_IS_SYS_DBNAME(p + 1)) {
if (p && IS_SYS_DBNAME(p + 1)) {
memmove(output->dbFName, p + 1, strlen(p + 1));
}
@ -2123,7 +2123,7 @@ int32_t ctgStartUpdateThread() {
int32_t ctgGetTbMetaFromCache(SCatalog* pCtg, SRequestConnInfo *pConn, SCtgTbMetaCtx* ctx, STableMeta** pTableMeta) {
if (CTG_IS_SYS_DBNAME(ctx->pName->dbname)) {
if (IS_SYS_DBNAME(ctx->pName->dbname)) {
CTG_FLAG_SET_SYS_DB(ctx->flag);
}
@ -2177,7 +2177,7 @@ _return:
}
int32_t ctgGetTbHashVgroupFromCache(SCatalog *pCtg, const SName *pTableName, SVgroupInfo **pVgroup) {
if (CTG_IS_SYS_DBNAME(pTableName->dbname)) {
if (IS_SYS_DBNAME(pTableName->dbname)) {
ctgError("no valid vgInfo for db, dbname:%s", pTableName->dbname);
CTG_ERR_RET(TSDB_CODE_CTG_INVALID_INPUT);
}

26
source/libs/catalog/src/ctgRemote.c
View File

@ -375,6 +375,8 @@ int32_t ctgGetQnodeListFromMnode(SCatalog* pCtg, SRequestConnInfo *pConn, SArray
CTG_ERR_RET(ctgProcessRspMsg(out, reqType, rpcRsp.pCont, rpcRsp.contLen, rpcRsp.code, NULL));
rpcFreeCont(rpcRsp.pCont);
return TSDB_CODE_SUCCESS;
}
@ -408,6 +410,8 @@ int32_t ctgGetDnodeListFromMnode(SCatalog* pCtg, SRequestConnInfo *pConn, SArray
CTG_ERR_RET(ctgProcessRspMsg(out, reqType, rpcRsp.pCont, rpcRsp.contLen, rpcRsp.code, NULL));
rpcFreeCont(rpcRsp.pCont);
return TSDB_CODE_SUCCESS;
}
@ -447,6 +451,8 @@ int32_t ctgGetDBVgInfoFromMnode(SCatalog* pCtg, SRequestConnInfo *pConn, SBuildU
CTG_ERR_RET(ctgProcessRspMsg(out, reqType, rpcRsp.pCont, rpcRsp.contLen, rpcRsp.code, input->db));
rpcFreeCont(rpcRsp.pCont);
return TSDB_CODE_SUCCESS;
}
@ -485,6 +491,8 @@ int32_t ctgGetDBCfgFromMnode(SCatalog* pCtg, SRequestConnInfo *pConn, const char
CTG_ERR_RET(ctgProcessRspMsg(out, reqType, rpcRsp.pCont, rpcRsp.contLen, rpcRsp.code, (char*)dbFName));
rpcFreeCont(rpcRsp.pCont);
return TSDB_CODE_SUCCESS;
}
@ -522,6 +530,8 @@ int32_t ctgGetIndexInfoFromMnode(SCatalog* pCtg, SRequestConnInfo *pConn, const
rpcSendRecv(pConn->pTrans, &pConn->mgmtEps, &rpcMsg, &rpcRsp);
CTG_ERR_RET(ctgProcessRspMsg(out, reqType, rpcRsp.pCont, rpcRsp.contLen, rpcRsp.code, (char*)indexName));
rpcFreeCont(rpcRsp.pCont);
return TSDB_CODE_SUCCESS;
}
@ -563,6 +573,8 @@ int32_t ctgGetTbIndexFromMnode(SCatalog* pCtg, SRequestConnInfo *pConn, SName *n
rpcSendRecv(pConn->pTrans, &pConn->mgmtEps, &rpcMsg, &rpcRsp);
CTG_ERR_RET(ctgProcessRspMsg(out, reqType, rpcRsp.pCont, rpcRsp.contLen, rpcRsp.code, (char*)tbFName));
rpcFreeCont(rpcRsp.pCont);
return TSDB_CODE_SUCCESS;
}
@ -602,6 +614,8 @@ int32_t ctgGetUdfInfoFromMnode(SCatalog* pCtg, SRequestConnInfo *pConn, const ch
CTG_ERR_RET(ctgProcessRspMsg(out, reqType, rpcRsp.pCont, rpcRsp.contLen, rpcRsp.code, (char*)funcName));
rpcFreeCont(rpcRsp.pCont);
return TSDB_CODE_SUCCESS;
}
@ -639,6 +653,8 @@ int32_t ctgGetUserDbAuthFromMnode(SCatalog* pCtg, SRequestConnInfo *pConn, const
rpcSendRecv(pConn->pTrans, &pConn->mgmtEps, &rpcMsg, &rpcRsp);
CTG_ERR_RET(ctgProcessRspMsg(out, reqType, rpcRsp.pCont, rpcRsp.contLen, rpcRsp.code, (char*)user));
rpcFreeCont(rpcRsp.pCont);
return TSDB_CODE_SUCCESS;
}
@ -683,6 +699,8 @@ int32_t ctgGetTbMetaFromMnodeImpl(SCatalog* pCtg, SRequestConnInfo *pConn, char
CTG_ERR_RET(ctgProcessRspMsg(out, reqType, rpcRsp.pCont, rpcRsp.contLen, rpcRsp.code, tbFName));
rpcFreeCont(rpcRsp.pCont);
return TSDB_CODE_SUCCESS;
}
@ -740,6 +758,8 @@ int32_t ctgGetTbMetaFromVnode(SCatalog* pCtg, SRequestConnInfo *pConn, const SNa
CTG_ERR_RET(ctgProcessRspMsg(out, reqType, rpcRsp.pCont, rpcRsp.contLen, rpcRsp.code, tbFName));
rpcFreeCont(rpcRsp.pCont);
return TSDB_CODE_SUCCESS;
}
@ -784,6 +804,8 @@ int32_t ctgGetTableCfgFromVnode(SCatalog* pCtg, SRequestConnInfo *pConn, const S
rpcSendRecv(pConn->pTrans, &vgroupInfo->epSet, &rpcMsg, &rpcRsp);
CTG_ERR_RET(ctgProcessRspMsg(out, reqType, rpcRsp.pCont, rpcRsp.contLen, rpcRsp.code, (char*)tbFName));
rpcFreeCont(rpcRsp.pCont);
return TSDB_CODE_SUCCESS;
}
@ -824,6 +846,8 @@ int32_t ctgGetTableCfgFromMnode(SCatalog* pCtg, SRequestConnInfo *pConn, const S
rpcSendRecv(pConn->pTrans, &pConn->mgmtEps, &rpcMsg, &rpcRsp);
CTG_ERR_RET(ctgProcessRspMsg(out, reqType, rpcRsp.pCont, rpcRsp.contLen, rpcRsp.code, (char*)tbFName));
rpcFreeCont(rpcRsp.pCont);
return TSDB_CODE_SUCCESS;
}
@ -858,6 +882,8 @@ int32_t ctgGetSvrVerFromMnode(SCatalog* pCtg, SRequestConnInfo *pConn, char **ou
rpcSendRecv(pConn->pTrans, &pConn->mgmtEps, &rpcMsg, &rpcRsp);
CTG_ERR_RET(ctgProcessRspMsg(out, reqType, rpcRsp.pCont, rpcRsp.contLen, rpcRsp.code, NULL));
rpcFreeCont(rpcRsp.pCont);
return TSDB_CODE_SUCCESS;
}

2
source/libs/command/src/explain.c
View File

@ -401,8 +401,6 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
QRY_ERR_RET(qExplainBufAppendExecInfo(pResNode->pExecInfo, tbuf, &tlen));
EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
}
EXPLAIN_ROW_APPEND(EXPLAIN_COLUMNS_FORMAT, pTagScanNode->pScanCols->length);
EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
if (pTagScanNode->pScanPseudoCols) {
EXPLAIN_ROW_APPEND(EXPLAIN_PSEUDO_COLUMNS_FORMAT, pTagScanNode->pScanPseudoCols->length);
EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);

3
source/libs/executor/inc/executorimpl.h
View File

@ -164,6 +164,7 @@ typedef struct {
char* dbname;
int32_t tversion;
SSchemaWrapper* sw;
SSchemaWrapper* qsw;
} SSchemaInfo;
typedef struct SExecTaskInfo {
@ -868,7 +869,7 @@ SOperatorInfo* createDataBlockInfoScanOperator(void* dataReader, SReadHandle* re
SExecTaskInfo* pTaskInfo);
SOperatorInfo* createStreamScanOperatorInfo(SReadHandle* pHandle, STableScanPhysiNode* pTableScanNode, SNode* pTagCond,
SExecTaskInfo* pTaskInfo, STimeWindowAggSupp* pTwSup);
STimeWindowAggSupp* pTwAggSup, SExecTaskInfo* pTaskInfo);
SOperatorInfo* createFillOperatorInfo(SOperatorInfo* downstream, SFillPhysiNode* pPhyFillNode, SExecTaskInfo* pTaskInfo);

10
source/libs/executor/src/executor.c
View File

@ -120,7 +120,8 @@ int32_t qSetMultiStreamInput(qTaskInfo_t tinfo, const void* pBlocks, size_t numO
return code;
}
qTaskInfo_t qCreateQueueExecTaskInfo(void* msg, SReadHandle* readers, int32_t* numOfCols) {
qTaskInfo_t qCreateQueueExecTaskInfo(void* msg, SReadHandle* readers, int32_t* numOfCols,
SSchemaWrapper** pSchemaWrapper) {
if (msg == NULL) {
// TODO create raw scan
return NULL;
@ -146,7 +147,7 @@ qTaskInfo_t qCreateQueueExecTaskInfo(void* msg, SReadHandle* readers, int32_t* n
SDataBlockDescNode* pDescNode = pPlan->pNode->pOutputDataBlockDesc;
*numOfCols = 0;
SNode* pNode;
SNode* pNode;
FOREACH(pNode, pDescNode->pSlots) {
SSlotDescNode* pSlotDesc = (SSlotDescNode*)pNode;
if (pSlotDesc->output) {
@ -154,6 +155,7 @@ qTaskInfo_t qCreateQueueExecTaskInfo(void* msg, SReadHandle* readers, int32_t* n
}
}
*pSchemaWrapper = tCloneSSchemaWrapper(((SExecTaskInfo*)pTaskInfo)->schemaInfo.qsw);
return pTaskInfo;
}
@ -248,9 +250,11 @@ int32_t qUpdateQualifiedTableId(qTaskInfo_t tinfo, const SArray* tableIdList, bo
// add to qTaskInfo
// todo refactor STableList
for(int32_t i = 0; i < taosArrayGetSize(qa); ++i) {
for (int32_t i = 0; i < taosArrayGetSize(qa); ++i) {
uint64_t* uid = taosArrayGet(qa, i);
qDebug("table %ld added to task info", *uid);
STableKeyInfo keyInfo = {.uid = *uid, .groupId = 0};
taosArrayPush(pTaskInfo->tableqinfoList.pTableList, &keyInfo);
}

26
source/libs/executor/src/executorMain.c
View File

@ -199,6 +199,10 @@ int32_t qAsyncKillTask(qTaskInfo_t qinfo) {
void qDestroyTask(qTaskInfo_t qTaskHandle) {
SExecTaskInfo* pTaskInfo = (SExecTaskInfo*)qTaskHandle;
if (pTaskInfo == NULL) {
return;
}
qDebug("%s execTask completed, numOfRows:%" PRId64, GET_TASKID(pTaskInfo), pTaskInfo->pRoot->resultInfo.totalRows);
queryCostStatis(pTaskInfo); // print the query cost summary
@ -311,6 +315,9 @@ int32_t qStreamPrepareScan(qTaskInfo_t tinfo, const STqOffsetVal* pOffset) {
if (type == QUERY_NODE_PHYSICAL_PLAN_STREAM_SCAN) {
SStreamScanInfo* pInfo = pOperator->info;
if (pOffset->type == TMQ_OFFSET__LOG) {
STableScanInfo* pTSInfo = pInfo->pTableScanOp->info;
tsdbReaderClose(pTSInfo->dataReader);
pTSInfo->dataReader = NULL;
#if 0
if (tOffsetEqual(pOffset, &pTaskInfo->streamInfo.lastStatus) &&
pInfo->tqReader->pWalReader->curVersion != pOffset->version) {
@ -337,11 +344,20 @@ int32_t qStreamPrepareScan(qTaskInfo_t tinfo, const STqOffsetVal* pOffset) {
return -1;
}
}
/*if (pTaskInfo->streamInfo.lastStatus.type != TMQ_OFFSET__SNAPSHOT_DATA ||*/
/*pTaskInfo->streamInfo.lastStatus.uid != uid || pTaskInfo->streamInfo.lastStatus.ts != ts) {*/
STableScanInfo* pTableScanInfo = pInfo->pTableScanOp->info;
int32_t tableSz = taosArrayGetSize(pTaskInfo->tableqinfoList.pTableList);
bool found = false;
#ifndef NDEBUG
qDebug("switch to next table %ld (cursor %d), %ld rows returned", uid, pTableScanInfo->currentTable,
pInfo->pTableScanOp->resultInfo.totalRows);
pInfo->pTableScanOp->resultInfo.totalRows = 0;
#endif
bool found = false;
for (int32_t i = 0; i < tableSz; i++) {
STableKeyInfo* pTableInfo = taosArrayGet(pTaskInfo->tableqinfoList.pTableList, i);
if (pTableInfo->uid == uid) {
@ -354,6 +370,14 @@ int32_t qStreamPrepareScan(qTaskInfo_t tinfo, const STqOffsetVal* pOffset) {
// TODO after dropping table, table may be not found
ASSERT(found);
if (pTableScanInfo->dataReader == NULL) {
if (tsdbReaderOpen(pTableScanInfo->readHandle.vnode, &pTableScanInfo->cond,
pTaskInfo->tableqinfoList.pTableList, &pTableScanInfo->dataReader, NULL) < 0 ||
pTableScanInfo->dataReader == NULL) {
ASSERT(0);
}
}
tsdbSetTableId(pTableScanInfo->dataReader, uid);
int64_t oldSkey = pTableScanInfo->cond.twindows.skey;
pTableScanInfo->cond.twindows.skey = ts + 1;

171
source/libs/executor/src/executorimpl.c
View File

@ -376,9 +376,7 @@ void initExecTimeWindowInfo(SColumnInfoData* pColData, STimeWindow* pQueryWindow
colDataAppendInt64(pColData, 4, &pQueryWindow->ekey);
}
void cleanupExecTimeWindowInfo(SColumnInfoData* pColData) {
colDataDestroy(pColData);
}
void cleanupExecTimeWindowInfo(SColumnInfoData* pColData) { colDataDestroy(pColData); }
void doApplyFunctions(SExecTaskInfo* taskInfo, SqlFunctionCtx* pCtx, STimeWindow* pWin,
SColumnInfoData* pTimeWindowData, int32_t offset, int32_t forwardStep, TSKEY* tsCol,
@ -524,8 +522,8 @@ static int32_t doSetInputDataBlock(SOperatorInfo* pOperator, SqlFunctionCtx* pCt
// NOTE: the last parameter is the primary timestamp column
// todo: refactor this
if (fmIsTimelineFunc(pCtx[i].functionId) && (j == pOneExpr->base.numOfParams - 1)) {
pInput->pPTS = pInput->pData[j]; // in case of merge function, this is not always the ts column data.
// ASSERT(pInput->pPTS->info.type == TSDB_DATA_TYPE_TIMESTAMP);
pInput->pPTS = pInput->pData[j]; // in case of merge function, this is not always the ts column data.
// ASSERT(pInput->pPTS->info.type == TSDB_DATA_TYPE_TIMESTAMP);
}
ASSERT(pInput->pData[j] != NULL);
} else if (pFuncParam->type == FUNC_PARAM_TYPE_VALUE) {
@ -633,7 +631,7 @@ int32_t projectApplyFunctions(SExprInfo* pExpr, SSDataBlock* pResult, SSDataBloc
ASSERT(pResult->info.capacity > 0);
colDataMergeCol(pResColData, startOffset, &pResult->info.capacity, &idata, dest.numOfRows);
colDataDestroy(&idata);
numOfRows = dest.numOfRows;
taosArrayDestroy(pBlockList);
} else if (pExpr[k].pExpr->nodeType == QUERY_NODE_FUNCTION) {
@ -835,7 +833,7 @@ void setTaskKilled(SExecTaskInfo* pTaskInfo) { pTaskInfo->code = TSDB_CODE_TSC_Q
/////////////////////////////////////////////////////////////////////////////////////////////
STimeWindow getAlignQueryTimeWindow(SInterval* pInterval, int32_t precision, int64_t key) {
STimeWindow win = {0};
STimeWindow win = {0};
win.skey = taosTimeTruncate(key, pInterval, precision);
/*
@ -1336,12 +1334,10 @@ void setResultRowInitCtx(SResultRow* pResult, SqlFunctionCtx* pCtx, int32_t numO
static void extractQualifiedTupleByFilterResult(SSDataBlock* pBlock, const int8_t* rowRes, bool keep);
void doFilter(const SNode* pFilterNode, SSDataBlock* pBlock) {
if (pFilterNode == NULL) {
return;
}
if (pBlock->info.rows == 0) {
if (pFilterNode == NULL || pBlock->info.rows == 0) {
return;
}
SFilterInfo* filter = NULL;
// todo move to the initialization function
@ -1358,8 +1354,6 @@ void doFilter(const SNode* pFilterNode, SSDataBlock* pBlock) {
filterFreeInfo(filter);
extractQualifiedTupleByFilterResult(pBlock, rowRes, keep);
blockDataUpdateTsWindow(pBlock, 0);
taosMemoryFree(rowRes);
}
@ -1647,11 +1641,6 @@ static int32_t compressQueryColData(SColumnInfoData* pColRes, int32_t numOfRows,
colSize + COMP_OVERFLOW_BYTES, compressed, NULL, 0);
}
int32_t doFillTimeIntervalGapsInResults(struct SFillInfo* pFillInfo, SSDataBlock* pBlock, int32_t capacity) {
int32_t numOfRows = (int32_t)taosFillResultDataBlock(pFillInfo, pBlock, capacity - pBlock->info.rows);
return pBlock->info.rows;
}
void queryCostStatis(SExecTaskInfo* pTaskInfo) {
STaskCostInfo* pSummary = &pTaskInfo->cost;
@ -2383,7 +2372,7 @@ static SSDataBlock* doLoadRemoteData(SOperatorInfo* pOperator) {
return NULL;
}
while(1) {
while (1) {
SSDataBlock* pBlock = doLoadRemoteDataImpl(pOperator);
if (pBlock == NULL) {
return NULL;
@ -2879,7 +2868,7 @@ int32_t getTableScanInfo(SOperatorInfo* pOperator, int32_t* order, int32_t* scan
*order = TSDB_ORDER_ASC;
*scanFlag = MAIN_SCAN;
return TSDB_CODE_SUCCESS;
} else if (type == QUERY_NODE_PHYSICAL_PLAN_TABLE_SCAN) {
} else if (type == QUERY_NODE_PHYSICAL_PLAN_TABLE_SCAN || type == QUERY_NODE_PHYSICAL_PLAN_TABLE_MERGE_SCAN) {
STableScanInfo* pTableScanInfo = pOperator->info;
*order = pTableScanInfo->cond.order;
*scanFlag = pTableScanInfo->scanFlag;
@ -3419,6 +3408,7 @@ static SSDataBlock* doFillImpl(SOperatorInfo* pOperator) {
doHandleRemainBlockFromNewGroup(pInfo, pResultInfo, pTaskInfo);
if (pResBlock->info.rows > pResultInfo->threshold || pResBlock->info.rows > 0) {
pResBlock->info.groupId = pInfo->curGroupId;
return pResBlock;
}
@ -3436,13 +3426,13 @@ static SSDataBlock* doFillImpl(SOperatorInfo* pOperator) {
blockDataUpdateTsWindow(pBlock, pInfo->primaryTsCol);
if (pInfo->curGroupId == 0 || pInfo->curGroupId == pBlock->info.groupId) {
pInfo->curGroupId = pBlock->info.groupId; // the first data block
pInfo->curGroupId = pBlock->info.groupId; // the first data block
pInfo->totalInputRows += pBlock->info.rows;
taosFillSetStartInfo(pInfo->pFillInfo, pBlock->info.rows, pBlock->info.window.ekey);
taosFillSetInputDataBlock(pInfo->pFillInfo, pBlock);
} else if (pInfo->curGroupId != pBlock->info.groupId) { // the new group data block
} else if (pInfo->curGroupId != pBlock->info.groupId) { // the new group data block
pInfo->existNewGroupBlock = pBlock;
// Fill the previous group data block, before handle the data block of new group.
@ -3461,17 +3451,20 @@ static SSDataBlock* doFillImpl(SOperatorInfo* pOperator) {
// 1. The result in current group not reach the threshold of output result, continue
// 2. If multiple group results existing in one SSDataBlock is not allowed, return immediately
if (pResBlock->info.rows > pResultInfo->threshold || pBlock == NULL || pInfo->existNewGroupBlock != NULL) {
pResBlock->info.groupId = pInfo->curGroupId;
return pResBlock;
}
doHandleRemainBlockFromNewGroup(pInfo, pResultInfo, pTaskInfo);
if (pResBlock->info.rows >= pOperator->resultInfo.threshold || pBlock == NULL) {
pResBlock->info.groupId = pInfo->curGroupId;
return pResBlock;
}
} else if (pInfo->existNewGroupBlock) { // try next group
assert(pBlock != NULL);
doHandleRemainBlockForNewGroupImpl(pInfo, pResultInfo, pTaskInfo);
if (pResBlock->info.rows > pResultInfo->threshold) {
pResBlock->info.groupId = pInfo->curGroupId;
return pResBlock;
}
} else {
@ -3491,23 +3484,19 @@ static SSDataBlock* doFill(SOperatorInfo* pOperator) {
SSDataBlock* fillResult = NULL;
while (true) {
fillResult = doFillImpl(pOperator);
if (fillResult != NULL) {
doFilter(pInfo->pCondition, fillResult);
}
if (fillResult == NULL) {
doSetOperatorCompleted(pOperator);
break;
}
doFilter(pInfo->pCondition, fillResult);
if (fillResult->info.rows > 0) {
break;
}
}
if (fillResult != NULL) {
size_t rows = fillResult->info.rows;
pOperator->resultInfo.totalRows += rows;
pOperator->resultInfo.totalRows += fillResult->info.rows;
}
return fillResult;
@ -3516,7 +3505,7 @@ static SSDataBlock* doFill(SOperatorInfo* pOperator) {
static void destroyExprInfo(SExprInfo* pExpr, int32_t numOfExprs) {
for (int32_t i = 0; i < numOfExprs; ++i) {
SExprInfo* pExprInfo = &pExpr[i];
for(int32_t j = 0; j < pExprInfo->base.numOfParams; ++j) {
for (int32_t j = 0; j < pExprInfo->base.numOfParams; ++j) {
if (pExprInfo->base.pParam[j].type == FUNC_PARAM_TYPE_COLUMN) {
taosMemoryFreeClear(pExprInfo->base.pParam[j].pCol);
}
@ -3609,7 +3598,7 @@ int32_t initAggInfo(SExprSupp* pSup, SAggSupporter* pAggSup, SExprInfo* pExprInf
return TSDB_CODE_SUCCESS;
}
void initResultSizeInfo(SResultInfo * pResultInfo, int32_t numOfRows) {
void initResultSizeInfo(SResultInfo* pResultInfo, int32_t numOfRows) {
ASSERT(numOfRows != 0);
pResultInfo->capacity = numOfRows;
pResultInfo->threshold = numOfRows * 0.75;
@ -3729,7 +3718,6 @@ void destroyBasicOperatorInfo(void* param, int32_t numOfOutput) {
taosMemoryFreeClear(param);
}
static void freeItem(void* pItem) {
void** p = pItem;
if (*p != NULL) {
@ -3963,6 +3951,8 @@ static SSDataBlock* doApplyIndefinitFunction(SOperatorInfo* pOperator) {
size_t rows = pInfo->pRes->info.rows;
if (rows > 0 || pOperator->status == OP_EXEC_DONE) {
break;
} else {
blockDataCleanup(pInfo->pRes);
}
}
@ -4054,8 +4044,8 @@ static int32_t initFillInfo(SFillOperatorInfo* pInfo, SExprInfo* pExpr, int32_t
w = getFirstQualifiedTimeWindow(win.skey, &w, pInterval, TSDB_ORDER_ASC);
int32_t order = TSDB_ORDER_ASC;
pInfo->pFillInfo = taosCreateFillInfo(order, w.skey, 0, capacity, numOfCols, pInterval,
fillType, pColInfo, pInfo->primaryTsCol, id);
pInfo->pFillInfo =
taosCreateFillInfo(order, w.skey, 0, capacity, numOfCols, pInterval, fillType, pColInfo, pInfo->primaryTsCol, id);
pInfo->win = win;
pInfo->p = taosMemoryCalloc(numOfCols, POINTER_BYTES);
@ -4069,7 +4059,8 @@ static int32_t initFillInfo(SFillOperatorInfo* pInfo, SExprInfo* pExpr, int32_t
}
}
SOperatorInfo* createFillOperatorInfo(SOperatorInfo* downstream, SFillPhysiNode* pPhyFillNode, SExecTaskInfo* pTaskInfo) {
SOperatorInfo* createFillOperatorInfo(SOperatorInfo* downstream, SFillPhysiNode* pPhyFillNode,
SExecTaskInfo* pTaskInfo) {
SFillOperatorInfo* pInfo = taosMemoryCalloc(1, sizeof(SFillOperatorInfo));
SOperatorInfo* pOperator = taosMemoryCalloc(1, sizeof(SOperatorInfo));
if (pInfo == NULL || pOperator == NULL) {
@ -4145,35 +4136,62 @@ static STsdbReader* doCreateDataReader(STableScanPhysiNode* pTableScanNode, SRea
static SArray* extractColumnInfo(SNodeList* pNodeList);
int32_t extractTableSchemaInfo(SReadHandle* pHandle, uint64_t uid, SExecTaskInfo* pTaskInfo) {
SSchemaWrapper* extractQueriedColumnSchema(SScanPhysiNode* pScanNode);
int32_t extractTableSchemaInfo(SReadHandle* pHandle, SScanPhysiNode* pScanNode, SExecTaskInfo* pTaskInfo) {
SMetaReader mr = {0};
metaReaderInit(&mr, pHandle->meta, 0);
int32_t code = metaGetTableEntryByUid(&mr, uid);
int32_t code = metaGetTableEntryByUid(&mr, pScanNode->uid);
if (code != TSDB_CODE_SUCCESS) {
qError("failed to get the table meta, uid:0x%" PRIx64 ", suid:0x%" PRIx64 ", %s", pScanNode->uid, pScanNode->suid,
GET_TASKID(pTaskInfo));
metaReaderClear(&mr);
return terrno;
}
pTaskInfo->schemaInfo.tablename = strdup(mr.me.name);
SSchemaInfo* pSchemaInfo = &pTaskInfo->schemaInfo;
pSchemaInfo->tablename = strdup(mr.me.name);
if (mr.me.type == TSDB_SUPER_TABLE) {
pTaskInfo->schemaInfo.sw = tCloneSSchemaWrapper(&mr.me.stbEntry.schemaRow);
pTaskInfo->schemaInfo.tversion = mr.me.stbEntry.schemaTag.version;
pSchemaInfo->sw = tCloneSSchemaWrapper(&mr.me.stbEntry.schemaRow);
pSchemaInfo->tversion = mr.me.stbEntry.schemaTag.version;
} else if (mr.me.type == TSDB_CHILD_TABLE) {
tDecoderClear(&mr.coder);
tb_uid_t suid = mr.me.ctbEntry.suid;
metaGetTableEntryByUid(&mr, suid);
pTaskInfo->schemaInfo.sw = tCloneSSchemaWrapper(&mr.me.stbEntry.schemaRow);
pTaskInfo->schemaInfo.tversion = mr.me.stbEntry.schemaTag.version;
pSchemaInfo->sw = tCloneSSchemaWrapper(&mr.me.stbEntry.schemaRow);
pSchemaInfo->tversion = mr.me.stbEntry.schemaTag.version;
} else {
pTaskInfo->schemaInfo.sw = tCloneSSchemaWrapper(&mr.me.ntbEntry.schemaRow);
pSchemaInfo->sw = tCloneSSchemaWrapper(&mr.me.ntbEntry.schemaRow);
}
metaReaderClear(&mr);
pSchemaInfo->qsw = extractQueriedColumnSchema(pScanNode);
return TSDB_CODE_SUCCESS;
}
SSchemaWrapper* extractQueriedColumnSchema(SScanPhysiNode* pScanNode) {
int32_t numOfCols = LIST_LENGTH(pScanNode->pScanCols);
SSchemaWrapper* pqSw = taosMemoryCalloc(1, sizeof(SSchemaWrapper));
pqSw->pSchema = taosMemoryCalloc(numOfCols, sizeof(SSchema));
for (int32_t i = 0; i < numOfCols; ++i) {
STargetNode* pNode = (STargetNode*)nodesListGetNode(pScanNode->pScanCols, i);
SColumnNode* pColNode = (SColumnNode*)pNode->pExpr;
SSchema* pSchema = &pqSw->pSchema[pqSw->nCols++];
pSchema->colId = pColNode->colId;
pSchema->type = pColNode->node.resType.type;
pSchema->type = pColNode->node.resType.bytes;
strncpy(pSchema->name, pColNode->colName, tListLen(pSchema->name));
}
return pqSw;
}
static void cleanupTableSchemaInfo(SSchemaInfo* pSchemaInfo) {
taosMemoryFreeClear(pSchemaInfo->dbname);
if (pSchemaInfo->sw == NULL) {
@ -4181,8 +4199,8 @@ static void cleanupTableSchemaInfo(SSchemaInfo* pSchemaInfo) {
}
taosMemoryFree(pSchemaInfo->tablename);
taosMemoryFree(pSchemaInfo->sw->pSchema);
taosMemoryFree(pSchemaInfo->sw);
tDeleteSSchemaWrapper(pSchemaInfo->sw);
tDeleteSSchemaWrapper(pSchemaInfo->qsw);
}
static int32_t sortTableGroup(STableListInfo* pTableListInfo, int32_t groupNum) {
@ -4291,7 +4309,7 @@ int32_t generateGroupIdMap(STableListInfo* pTableListInfo, SReadHandle* pHandle,
REPLACE_NODE(pNew);
} else {
taosMemoryFree(keyBuf);
nodesClearList(groupNew);
nodesDestroyList(groupNew);
metaReaderClear(&mr);
return code;
}
@ -4309,7 +4327,7 @@ int32_t generateGroupIdMap(STableListInfo* pTableListInfo, SReadHandle* pHandle,
if (tTagIsJson(data)) {
terrno = TSDB_CODE_QRY_JSON_IN_GROUP_ERROR;
taosMemoryFree(keyBuf);
nodesClearList(groupNew);
nodesDestroyList(groupNew);
metaReaderClear(&mr);
return terrno;
}
@ -4332,7 +4350,7 @@ int32_t generateGroupIdMap(STableListInfo* pTableListInfo, SReadHandle* pHandle,
info->groupId = groupId;
groupNum++;
nodesClearList(groupNew);
nodesDestroyList(groupNew);
metaReaderClear(&mr);
}
taosMemoryFree(keyBuf);
@ -4363,27 +4381,29 @@ static int32_t initTableblockDistQueryCond(uint64_t uid, SQueryTableDataCond* pC
pCond->suid = uid;
pCond->type = BLOCK_LOAD_OFFSET_ORDER;
pCond->startVersion = -1;
pCond->endVersion = -1;
pCond->endVersion = -1;
return TSDB_CODE_SUCCESS;
}
SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo, SReadHandle* pHandle,
STableListInfo* pTableListInfo, SNode* pTagCond, SNode* pTagIndexCond, const char* pUser) {
STableListInfo* pTableListInfo, SNode* pTagCond, SNode* pTagIndexCond,
const char* pUser) {
int32_t type = nodeType(pPhyNode);
if (pPhyNode->pChildren == NULL || LIST_LENGTH(pPhyNode->pChildren) == 0) {
if (QUERY_NODE_PHYSICAL_PLAN_TABLE_SCAN == type) {
STableScanPhysiNode* pTableScanNode = (STableScanPhysiNode*)pPhyNode;
int32_t code = createScanTableListInfo(&pTableScanNode->scan, pTableScanNode->pGroupTags,
pTableScanNode->groupSort, pHandle, pTableListInfo, pTagCond, pTagIndexCond, GET_TASKID(pTaskInfo));
int32_t code =
createScanTableListInfo(&pTableScanNode->scan, pTableScanNode->pGroupTags, pTableScanNode->groupSort, pHandle,
pTableListInfo, pTagCond, pTagIndexCond, GET_TASKID(pTaskInfo));
if (code) {
pTaskInfo->code = code;
return NULL;
}
code = extractTableSchemaInfo(pHandle, pTableScanNode->scan.uid, pTaskInfo);
code = extractTableSchemaInfo(pHandle, &pTableScanNode->scan, pTaskInfo);
if (code) {
pTaskInfo->code = terrno;
return NULL;
@ -4396,21 +4416,21 @@ SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo
} else if (QUERY_NODE_PHYSICAL_PLAN_TABLE_MERGE_SCAN == type) {
STableMergeScanPhysiNode* pTableScanNode = (STableMergeScanPhysiNode*)pPhyNode;
int32_t code = createScanTableListInfo(&pTableScanNode->scan, pTableScanNode->pGroupTags,
pTableScanNode->groupSort, pHandle, pTableListInfo, pTagCond, pTagIndexCond, GET_TASKID(pTaskInfo));
int32_t code =
createScanTableListInfo(&pTableScanNode->scan, pTableScanNode->pGroupTags, pTableScanNode->groupSort, pHandle,
pTableListInfo, pTagCond, pTagIndexCond, GET_TASKID(pTaskInfo));
if (code) {
pTaskInfo->code = code;
return NULL;
}
code = extractTableSchemaInfo(pHandle, pTableScanNode->scan.uid, pTaskInfo);
code = extractTableSchemaInfo(pHandle, &pTableScanNode->scan, pTaskInfo);
if (code) {
pTaskInfo->code = terrno;
return NULL;
}
SOperatorInfo* pOperator =
createTableMergeScanOperatorInfo(pTableScanNode, pTableListInfo, pHandle, pTaskInfo);
SOperatorInfo* pOperator = createTableMergeScanOperatorInfo(pTableScanNode, pTableListInfo, pHandle, pTaskInfo);
STableScanInfo* pScanInfo = pOperator->info;
pTaskInfo->cost.pRecoder = &pScanInfo->readRecorder;
@ -4420,21 +4440,32 @@ SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo
return createExchangeOperatorInfo(pHandle->pMsgCb->clientRpc, (SExchangePhysiNode*)pPhyNode, pTaskInfo);
} else if (QUERY_NODE_PHYSICAL_PLAN_STREAM_SCAN == type) {
STableScanPhysiNode* pTableScanNode = (STableScanPhysiNode*)pPhyNode;
STimeWindowAggSupp twSup = {
.waterMark = pTableScanNode->watermark,
.calTrigger = pTableScanNode->triggerType,
.maxTs = INT64_MIN,
STimeWindowAggSupp aggSup = (STimeWindowAggSupp){
.waterMark = pTableScanNode->watermark,
.calTrigger = pTableScanNode->triggerType,
.maxTs = INT64_MIN,
};
if (pHandle->vnode) {
int32_t code = createScanTableListInfo(&pTableScanNode->scan, pTableScanNode->pGroupTags,
pTableScanNode->groupSort, pHandle, pTableListInfo, pTagCond, pTagIndexCond, GET_TASKID(pTaskInfo));
int32_t code =
createScanTableListInfo(&pTableScanNode->scan, pTableScanNode->pGroupTags, pTableScanNode->groupSort,
pHandle, pTableListInfo, pTagCond, pTagIndexCond, GET_TASKID(pTaskInfo));
if (code) {
pTaskInfo->code = code;
return NULL;
}
}
SOperatorInfo* pOperator = createStreamScanOperatorInfo(pHandle, pTableScanNode, pTagCond, pTaskInfo, &twSup);
#ifndef NDEBUG
int32_t sz = taosArrayGetSize(pTableListInfo->pTableList);
for (int32_t i = 0; i < sz; i++) {
STableKeyInfo* pKeyInfo = taosArrayGet(pTableListInfo->pTableList, i);
qDebug("creating stream task: add table %ld", pKeyInfo->uid);
}
}
#endif
pTaskInfo->schemaInfo.qsw = extractQueriedColumnSchema(&pTableScanNode->scan);
SOperatorInfo* pOperator = createStreamScanOperatorInfo(pHandle, pTableScanNode, pTagCond, &aggSup, pTaskInfo);
return pOperator;
} else if (QUERY_NODE_PHYSICAL_PLAN_SYSTABLE_SCAN == type) {
@ -4466,7 +4497,7 @@ SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo
}
SQueryTableDataCond cond = {0};
int32_t code = initTableblockDistQueryCond(pBlockNode->suid, &cond);
int32_t code = initTableblockDistQueryCond(pBlockNode->suid, &cond);
if (code != TSDB_CODE_SUCCESS) {
return NULL;
}
@ -4479,13 +4510,14 @@ SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo
} else if (QUERY_NODE_PHYSICAL_PLAN_LAST_ROW_SCAN == type) {
SLastRowScanPhysiNode* pScanNode = (SLastRowScanPhysiNode*)pPhyNode;
int32_t code = createScanTableListInfo(&pScanNode->scan, pScanNode->pGroupTags, true, pHandle, pTableListInfo, pTagCond, pTagIndexCond, GET_TASKID(pTaskInfo));
int32_t code = createScanTableListInfo(&pScanNode->scan, pScanNode->pGroupTags, true, pHandle, pTableListInfo,
pTagCond, pTagIndexCond, GET_TASKID(pTaskInfo));
if (code != TSDB_CODE_SUCCESS) {
pTaskInfo->code = code;
return NULL;
}
code = extractTableSchemaInfo(pHandle, pScanNode->scan.uid, pTaskInfo);
code = extractTableSchemaInfo(pHandle, &pScanNode->scan, pTaskInfo);
if (code != TSDB_CODE_SUCCESS) {
pTaskInfo->code = code;
return NULL;
@ -4941,7 +4973,8 @@ int32_t createExecTaskInfoImpl(SSubplan* pPlan, SExecTaskInfo** pTaskInfo, SRead
(*pTaskInfo)->sql = sql;
(*pTaskInfo)->pSubplan = pPlan;
(*pTaskInfo)->pRoot = createOperatorTree(pPlan->pNode, *pTaskInfo, pHandle, &(*pTaskInfo)->tableqinfoList, pPlan->pTagCond, pPlan->pTagIndexCond, pPlan->user);
(*pTaskInfo)->pRoot = createOperatorTree(pPlan->pNode, *pTaskInfo, pHandle, &(*pTaskInfo)->tableqinfoList,
pPlan->pTagCond, pPlan->pTagIndexCond, pPlan->user);
if (NULL == (*pTaskInfo)->pRoot) {
code = (*pTaskInfo)->code;

49
source/libs/executor/src/scanoperator.c
View File

@ -359,6 +359,7 @@ void setTbNameColData(void* pMeta, const SSDataBlock* pBlock, SColumnInfoData* p
SScalarParam param = {.columnData = pColInfoData};
fpSet.process(&srcParam, 1, &param);
colDataDestroy(&infoData);
}
static SSDataBlock* doTableScanImpl(SOperatorInfo* pOperator) {
@ -739,7 +740,7 @@ static SSDataBlock* doBlockInfoScan(SOperatorInfo* pOperator) {
static void destroyBlockDistScanOperatorInfo(void* param, int32_t numOfOutput) {
SBlockDistInfo* pDistInfo = (SBlockDistInfo*)param;
blockDataDestroy(pDistInfo->pResBlock);
tsdbReaderClose(pDistInfo->pHandle);
taosMemoryFreeClear(param);
}
@ -981,6 +982,9 @@ static SSDataBlock* doRangeScan(SStreamScanInfo* pInfo, SSDataBlock* pSDB, int32
if (!pResult) {
blockDataCleanup(pSDB);
*pRowIndex = 0;
STableScanInfo* pTableScanInfo = pInfo->pTableScanOp->info;
tsdbReaderClose(pTableScanInfo->dataReader);
pTableScanInfo->dataReader = NULL;
return NULL;
}
@ -1002,6 +1006,9 @@ static SSDataBlock* doDataScan(SStreamScanInfo* pInfo, SSDataBlock* pSDB, int32_
}
if (!pResult) {
pInfo->updateWin = (STimeWindow){.skey = INT64_MIN, .ekey = INT64_MAX};
STableScanInfo* pTableScanInfo = pInfo->pTableScanOp->info;
tsdbReaderClose(pTableScanInfo->dataReader);
pTableScanInfo->dataReader = NULL;
return NULL;
}
@ -1524,7 +1531,7 @@ static void destroyStreamScanOperatorInfo(void* param, int32_t numOfOutput) {
}
SOperatorInfo* createStreamScanOperatorInfo(SReadHandle* pHandle, STableScanPhysiNode* pTableScanNode, SNode* pTagCond,
SExecTaskInfo* pTaskInfo, STimeWindowAggSupp* pTwSup) {
STimeWindowAggSupp* pTwSup, SExecTaskInfo* pTaskInfo) {
SStreamScanInfo* pInfo = taosMemoryCalloc(1, sizeof(SStreamScanInfo));
SOperatorInfo* pOperator = taosMemoryCalloc(1, sizeof(SOperatorInfo));
@ -1538,6 +1545,8 @@ SOperatorInfo* createStreamScanOperatorInfo(SReadHandle* pHandle, STableScanPhys
pInfo->pTagCond = pTagCond;
pInfo->twAggSup = *pTwSup;
int32_t numOfCols = 0;
pInfo->pColMatchInfo = extractColMatchInfo(pScanPhyNode->pScanCols, pDescNode, &numOfCols, COL_MATCH_FROM_COL_ID);
@ -1590,7 +1599,7 @@ SOperatorInfo* createStreamScanOperatorInfo(SReadHandle* pHandle, STableScanPhys
}
if (pTSInfo->interval.interval > 0) {
pInfo->pUpdateInfo = updateInfoInitP(&pTSInfo->interval, pTwSup->waterMark);
pInfo->pUpdateInfo = updateInfoInitP(&pTSInfo->interval, pInfo->twAggSup.waterMark);
} else {
pInfo->pUpdateInfo = NULL;
}
@ -1630,7 +1639,6 @@ SOperatorInfo* createStreamScanOperatorInfo(SReadHandle* pHandle, STableScanPhys
pInfo->deleteDataIndex = 0;
pInfo->pDeleteDataRes = createPullDataBlock();
pInfo->updateWin = (STimeWindow){.skey = INT64_MAX, .ekey = INT64_MAX};
pInfo->twAggSup = *pTwSup;
pOperator->name = "StreamScanOperator";
pOperator->operatorType = QUERY_NODE_PHYSICAL_PLAN_STREAM_SCAN;
@ -2045,8 +2053,8 @@ static SSDataBlock* sysTableScanUserTables(SOperatorInfo* pOperator) {
uint64_t suid = pInfo->pCur->mr.me.ctbEntry.suid;
int32_t code = metaGetTableEntryByUid(&mr, suid);
if (code != TSDB_CODE_SUCCESS) {
qError("failed to get super table meta, uid:0x%" PRIx64 ", code:%s, %s", suid, tstrerror(terrno),
GET_TASKID(pTaskInfo));
qError("failed to get super table meta, cname:%s, suid:0x%" PRIx64 ", code:%s, %s", pInfo->pCur->mr.me.name,
suid, tstrerror(terrno), GET_TASKID(pTaskInfo));
metaReaderClear(&mr);
metaCloseTbCursor(pInfo->pCur);
pInfo->pCur = NULL;
@ -2152,16 +2160,39 @@ static SSDataBlock* sysTableScanUserTables(SOperatorInfo* pOperator) {
}
}
static SSDataBlock* sysTableScanUserSTables(SOperatorInfo* pOperator) {
SExecTaskInfo* pTaskInfo = pOperator->pTaskInfo;
SSysTableScanInfo* pInfo = pOperator->info;
if (pOperator->status == OP_EXEC_DONE) {
return NULL;
}
pInfo->pRes->info.rows = 0;
pOperator->status = OP_EXEC_DONE;
pInfo->loadInfo.totalRows += pInfo->pRes->info.rows;
return (pInfo->pRes->info.rows == 0) ? NULL : pInfo->pRes;
}
static SSDataBlock* doSysTableScan(SOperatorInfo* pOperator) {
// build message and send to mnode to fetch the content of system tables.
SExecTaskInfo* pTaskInfo = pOperator->pTaskInfo;
SSysTableScanInfo* pInfo = pOperator->info;
const char* name = tNameGetTableName(&pInfo->name);
if (pInfo->showRewrite) {
char dbName[TSDB_DB_NAME_LEN] = {0};
getDBNameFromCondition(pInfo->pCondition, dbName);
sprintf(pInfo->req.db, "%d.%s", pInfo->accountId, dbName);
}
if (strncasecmp(name, TSDB_INS_TABLE_USER_TABLES, TSDB_TABLE_FNAME_LEN) == 0) {
return sysTableScanUserTables(pOperator);
} else if (strncasecmp(name, TSDB_INS_TABLE_USER_TAGS, TSDB_TABLE_FNAME_LEN) == 0) {
return sysTableScanUserTags(pOperator);
} else if (strncasecmp(name, TSDB_INS_TABLE_USER_STABLES, TSDB_TABLE_FNAME_LEN) == 0 &&
IS_SYS_DBNAME(pInfo->req.db)) {
return sysTableScanUserSTables(pOperator);
} else { // load the meta from mnode of the given epset
if (pOperator->status == OP_EXEC_DONE) {
return NULL;
@ -2172,12 +2203,6 @@ static SSDataBlock* doSysTableScan(SOperatorInfo* pOperator) {
strncpy(pInfo->req.tb, tNameGetTableName(&pInfo->name), tListLen(pInfo->req.tb));
strcpy(pInfo->req.user, pInfo->pUser);
if (pInfo->showRewrite) {
char dbName[TSDB_DB_NAME_LEN] = {0};
getDBNameFromCondition(pInfo->pCondition, dbName);
sprintf(pInfo->req.db, "%d.%s", pInfo->accountId, dbName);
}
int32_t contLen = tSerializeSRetrieveTableReq(NULL, 0, &pInfo->req);
char* buf1 = taosMemoryCalloc(1, contLen);
tSerializeSRetrieveTableReq(buf1, contLen, &pInfo->req);

86
source/libs/executor/src/tfill.c
View File

@ -66,12 +66,32 @@ static void setNullRow(SSDataBlock* pBlock, int64_t ts, int32_t rowIndex) {
static void doSetVal(SColumnInfoData* pDstColInfoData, int32_t rowIndex, const SGroupKeys* pKey);
static void doSetUserSpecifiedValue(SColumnInfoData* pDst, SVariant* pVar, int32_t rowIndex, int64_t currentKey) {
if (pDst->info.type == TSDB_DATA_TYPE_FLOAT) {
float v = 0;
GET_TYPED_DATA(v, float, pVar->nType, &pVar->i);
colDataAppend(pDst, rowIndex, (char*)&v, false);
} else if (pDst->info.type == TSDB_DATA_TYPE_DOUBLE) {
double v = 0;
GET_TYPED_DATA(v, double, pVar->nType, &pVar->i);
colDataAppend(pDst, rowIndex, (char*)&v, false);
} else if (IS_SIGNED_NUMERIC_TYPE(pDst->info.type)) {
int64_t v = 0;
GET_TYPED_DATA(v, int64_t, pVar->nType, &pVar->i);
colDataAppend(pDst, rowIndex, (char*)&v, false);
} else if (pDst->info.type == TSDB_DATA_TYPE_TIMESTAMP) {
colDataAppend(pDst, rowIndex, (const char*)&currentKey, false);
} else { // varchar/nchar data
colDataAppendNULL(pDst, rowIndex);
}
}
static void doFillOneRow(SFillInfo* pFillInfo, SSDataBlock* pBlock, SSDataBlock* pSrcBlock, int64_t ts,
bool outOfBound) {
SPoint point1, point2, point;
int32_t step = GET_FORWARD_DIRECTION_FACTOR(pFillInfo->order);
// set the primary timestamp column value
// set the primary timestamp column value
int32_t index = pBlock->info.rows;
// set the other values
@ -160,30 +180,13 @@ static void doFillOneRow(SFillInfo* pFillInfo, SSDataBlock* pBlock, SSDataBlock*
} else { // fill with user specified value for each column
for (int32_t i = 0; i < pFillInfo->numOfCols; ++i) {
SFillColInfo* pCol = &pFillInfo->pFillCol[i];
if (TSDB_COL_IS_TAG(pCol->flag) /* || IS_VAR_DATA_TYPE(pCol->schema.type)*/) {
if (TSDB_COL_IS_TAG(pCol->flag)) {
continue;
}
SVariant* pVar = &pFillInfo->pFillCol[i].fillVal;
SColumnInfoData* pDst = taosArrayGet(pBlock->pDataBlock, i);
if (pDst->info.type == TSDB_DATA_TYPE_FLOAT) {
float v = 0;
GET_TYPED_DATA(v, float, pVar->nType, &pVar->i);
colDataAppend(pDst, index, (char*)&v, false);
} else if (pDst->info.type == TSDB_DATA_TYPE_DOUBLE) {
double v = 0;
GET_TYPED_DATA(v, double, pVar->nType, &pVar->i);
colDataAppend(pDst, index, (char*)&v, false);
} else if (IS_SIGNED_NUMERIC_TYPE(pDst->info.type)) {
int64_t v = 0;
GET_TYPED_DATA(v, int64_t, pVar->nType, &pVar->i);
colDataAppend(pDst, index, (char*)&v, false);
} else if (pDst->info.type == TSDB_DATA_TYPE_TIMESTAMP) {
colDataAppend(pDst, index, (const char*)&pFillInfo->currentKey, false);
} else { // varchar/nchar data
colDataAppendNULL(pDst, index);
}
doSetUserSpecifiedValue(pDst, pVar, index, pFillInfo->currentKey);
}
}
@ -273,7 +276,7 @@ static int32_t fillResultImpl(SFillInfo* pFillInfo, SSDataBlock* pBlock, int32_t
return outputRows;
}
} else {
assert(pFillInfo->currentKey == ts);
ASSERT(pFillInfo->currentKey == ts);
int32_t index = pBlock->info.rows;
if (pFillInfo->type == TSDB_FILL_NEXT && (pFillInfo->index + 1) < pFillInfo->numOfRows) {
@ -295,27 +298,32 @@ static int32_t fillResultImpl(SFillInfo* pFillInfo, SSDataBlock* pBlock, int32_t
SColumnInfoData* pSrc = taosArrayGet(pFillInfo->pSrcBlock->pDataBlock, srcSlotId);
char* src = colDataGetData(pSrc, pFillInfo->index);
if (i == 0 || (/*pCol->functionId != FUNCTION_COUNT &&*/ !colDataIsNull_s(pSrc, pFillInfo->index)) /*||
(pCol->functionId == FUNCTION_COUNT && GET_INT64_VAL(src) != 0)*/) {
if (/*i == 0 || (*/!colDataIsNull_s(pSrc, pFillInfo->index)) {
bool isNull = colDataIsNull_s(pSrc, pFillInfo->index);
colDataAppend(pDst, index, src, isNull);
saveColData(pFillInfo->prev, i, src, isNull);
} else { // i > 0 and data is null , do interpolation
if (pFillInfo->type == TSDB_FILL_PREV) {
SGroupKeys* pKey = taosArrayGet(pFillInfo->prev, i);
doSetVal(pDst, index, pKey);
} else if (pFillInfo->type == TSDB_FILL_LINEAR) {
bool isNull = colDataIsNull_s(pSrc, pFillInfo->index);
colDataAppend(pDst, index, src, isNull);
saveColData(pFillInfo->prev, i, src, isNull);
} else if (pFillInfo->type == TSDB_FILL_NULL) {
colDataAppendNULL(pDst, index);
} else if (pFillInfo->type == TSDB_FILL_NEXT) {
SGroupKeys* pKey = taosArrayGet(pFillInfo->next, i);
doSetVal(pDst, index, pKey);
} else {
SVariant* pVar = &pFillInfo->pFillCol[i].fillVal;
colDataAppend(pDst, index, (char*)&pVar->i, false);
} else {
if (pDst->info.type == TSDB_DATA_TYPE_TIMESTAMP) {
colDataAppend(pDst, index, (const char*)&pFillInfo->currentKey, false);
} else { // i > 0 and data is null , do interpolation
if (pFillInfo->type == TSDB_FILL_PREV) {
SArray* p = FILL_IS_ASC_FILL(pFillInfo) ? pFillInfo->prev : pFillInfo->next;
SGroupKeys* pKey = taosArrayGet(p, i);
doSetVal(pDst, index, pKey);
} else if (pFillInfo->type == TSDB_FILL_LINEAR) {
bool isNull = colDataIsNull_s(pSrc, pFillInfo->index);
colDataAppend(pDst, index, src, isNull);
saveColData(pFillInfo->prev, i, src, isNull); // todo:
} else if (pFillInfo->type == TSDB_FILL_NULL) {
colDataAppendNULL(pDst, index);
} else if (pFillInfo->type == TSDB_FILL_NEXT) {
SArray* p = FILL_IS_ASC_FILL(pFillInfo) ? pFillInfo->next : pFillInfo->prev;
SGroupKeys* pKey = taosArrayGet(p, i);
doSetVal(pDst, index, pKey);
} else {
SVariant* pVar = &pFillInfo->pFillCol[i].fillVal;
doSetUserSpecifiedValue(pDst, pVar, index, pFillInfo->currentKey);
}
}
}
}

1
source/libs/function/inc/functionMgtInt.h
View File

@ -47,6 +47,7 @@ extern "C" {
#define FUNC_MGT_SYSTEM_INFO_FUNC FUNC_MGT_FUNC_CLASSIFICATION_MASK(18)
#define FUNC_MGT_CLIENT_PC_FUNC FUNC_MGT_FUNC_CLASSIFICATION_MASK(19)
#define FUNC_MGT_MULTI_ROWS_FUNC FUNC_MGT_FUNC_CLASSIFICATION_MASK(20)
#define FUNC_MGT_KEEP_ORDER_FUNC FUNC_MGT_FUNC_CLASSIFICATION_MASK(21)
#define FUNC_MGT_TEST_MASK(val, mask) (((val) & (mask)) != 0)

10
source/libs/function/src/builtins.c
View File

@ -2088,7 +2088,7 @@ const SBuiltinFuncDefinition funcMgtBuiltins[] = {
.classification = FUNC_MGT_AGG_FUNC,
.translateFunc = translateApercentileMerge,
.getEnvFunc = getApercentileFuncEnv,
.initFunc = functionSetup,
.initFunc = apercentileFunctionSetup,
.processFunc = apercentileFunctionMerge,
.finalizeFunc = apercentileFinalize,
.invertFunc = NULL,
@ -2097,7 +2097,7 @@ const SBuiltinFuncDefinition funcMgtBuiltins[] = {
{
.name = "top",
.type = FUNCTION_TYPE_TOP,
.classification = FUNC_MGT_AGG_FUNC | FUNC_MGT_SELECT_FUNC | FUNC_MGT_MULTI_ROWS_FUNC | FUNC_MGT_FORBID_STREAM_FUNC,
.classification = FUNC_MGT_AGG_FUNC | FUNC_MGT_SELECT_FUNC | FUNC_MGT_MULTI_ROWS_FUNC | FUNC_MGT_KEEP_ORDER_FUNC | FUNC_MGT_FORBID_STREAM_FUNC,
.translateFunc = translateTopBot,
.getEnvFunc = getTopBotFuncEnv,
.initFunc = topBotFunctionSetup,
@ -2112,7 +2112,7 @@ const SBuiltinFuncDefinition funcMgtBuiltins[] = {
{
.name = "bottom",
.type = FUNCTION_TYPE_BOTTOM,
.classification = FUNC_MGT_AGG_FUNC | FUNC_MGT_SELECT_FUNC | FUNC_MGT_MULTI_ROWS_FUNC | FUNC_MGT_FORBID_STREAM_FUNC,
.classification = FUNC_MGT_AGG_FUNC | FUNC_MGT_SELECT_FUNC | FUNC_MGT_MULTI_ROWS_FUNC | FUNC_MGT_KEEP_ORDER_FUNC | FUNC_MGT_FORBID_STREAM_FUNC,
.translateFunc = translateTopBot,
.getEnvFunc = getTopBotFuncEnv,
.initFunc = topBotFunctionSetup,
@ -2480,7 +2480,7 @@ const SBuiltinFuncDefinition funcMgtBuiltins[] = {
{
.name = "sample",
.type = FUNCTION_TYPE_SAMPLE,
.classification = FUNC_MGT_AGG_FUNC | FUNC_MGT_SELECT_FUNC | FUNC_MGT_MULTI_ROWS_FUNC | FUNC_MGT_FORBID_STREAM_FUNC,
.classification = FUNC_MGT_AGG_FUNC | FUNC_MGT_SELECT_FUNC | FUNC_MGT_MULTI_ROWS_FUNC | FUNC_MGT_KEEP_ORDER_FUNC | FUNC_MGT_FORBID_STREAM_FUNC,
.translateFunc = translateSample,
.getEnvFunc = getSampleFuncEnv,
.initFunc = sampleFunctionSetup,
@ -2906,7 +2906,7 @@ const SBuiltinFuncDefinition funcMgtBuiltins[] = {
{
.name = "_select_value",
.type = FUNCTION_TYPE_SELECT_VALUE,
.classification = FUNC_MGT_AGG_FUNC | FUNC_MGT_SELECT_FUNC,
.classification = FUNC_MGT_AGG_FUNC | FUNC_MGT_SELECT_FUNC | FUNC_MGT_KEEP_ORDER_FUNC,
.translateFunc = translateSelectValue,
.getEnvFunc = getSelectivityFuncEnv, // todo remove this function later.
.initFunc = functionSetup,

2
source/libs/function/src/builtinsimpl.c
View File

@ -2592,6 +2592,7 @@ int32_t apercentileFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock) {
SAPercentileInfo* pInfo = (SAPercentileInfo*)GET_ROWCELL_INTERBUF(pResInfo);
if (pInfo->algo == APERCT_ALGO_TDIGEST) {
buildTDigestInfo(pInfo);
if (pInfo->pTDigest->size > 0) {
pInfo->result = tdigestQuantile(pInfo->pTDigest, pInfo->percent / 100);
} else { // no need to free
@ -2599,6 +2600,7 @@ int32_t apercentileFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock) {
return TSDB_CODE_SUCCESS;
}
} else {
buildHistogramInfo(pInfo);
if (pInfo->pHisto->numOfElems > 0) {
qDebug("get the final res:%d, elements:%"PRId64", entry:%d", pInfo->pHisto->numOfElems, pInfo->pHisto->numOfEntries);

2
source/libs/function/src/functionMgt.c
View File

@ -183,6 +183,8 @@ bool fmIsClientPseudoColumnFunc(int32_t funcId) { return isSpecificClassifyFunc(
bool fmIsMultiRowsFunc(int32_t funcId) { return isSpecificClassifyFunc(funcId, FUNC_MGT_MULTI_ROWS_FUNC); }
bool fmIsKeepOrderFunc(int32_t funcId) { return isSpecificClassifyFunc(funcId, FUNC_MGT_KEEP_ORDER_FUNC); }
bool fmIsInterpFunc(int32_t funcId) {
if (funcId < 0 || funcId >= funcMgtBuiltinsNum) {
return false;

4
source/libs/function/src/udfd.c
View File

@ -913,8 +913,8 @@ void udfdConnectMnodeThreadFunc(void *args) {
}
int main(int argc, char *argv[]) {
if (!taosCheckSystemIsSmallEnd()) {
printf("failed to start since on non-small-end machines\n");
if (!taosCheckSystemIsLittleEnd()) {
printf("failed to start since on non-little-end machines\n");
return -1;
}

2
source/libs/index/src/indexFilter.c
View File

@ -707,6 +707,8 @@ static int32_t sifCalculate(SNode *pNode, SIFParam *pDst) {
sifFreeParam(res);
taosHashRemove(ctx.pRes, (void *)&pNode, POINTER_BYTES);
}
sifFreeRes(ctx.pRes);
SIF_RET(code);
}

3
source/libs/nodes/src/nodesCloneFuncs.c
View File

@ -332,6 +332,9 @@ static int32_t logicNodeCopy(const SLogicNode* pSrc, SLogicNode* pDst) {
COPY_SCALAR_FIELD(precision);
CLONE_NODE_FIELD(pLimit);
CLONE_NODE_FIELD(pSlimit);
COPY_SCALAR_FIELD(requireDataOrder);
COPY_SCALAR_FIELD(resultDataOrder);
COPY_SCALAR_FIELD(groupAction);
return TSDB_CODE_SUCCESS;
}

21
source/libs/nodes/src/nodesCodeFuncs.c
View File

@ -504,6 +504,9 @@ static const char* jkLogicPlanConditions = "Conditions";
static const char* jkLogicPlanChildren = "Children";
static const char* jkLogicPlanLimit = "Limit";
static const char* jkLogicPlanSlimit = "SLimit";
static const char* jkLogicPlanRequireDataOrder = "RequireDataOrder";
static const char* jkLogicPlanResultDataOrder = "ResultDataOrder";
static const char* jkLogicPlanGroupAction = "GroupAction";
static int32_t logicPlanNodeToJson(const void* pObj, SJson* pJson) {
const SLogicNode* pNode = (const SLogicNode*)pObj;
@ -521,6 +524,15 @@ static int32_t logicPlanNodeToJson(const void* pObj, SJson* pJson) {
if (TSDB_CODE_SUCCESS == code) {
code = tjsonAddObject(pJson, jkLogicPlanSlimit, nodeToJson, pNode->pSlimit);
}
if (TSDB_CODE_SUCCESS == code) {
code = tjsonAddIntegerToObject(pJson, jkLogicPlanRequireDataOrder, pNode->requireDataOrder);
}
if (TSDB_CODE_SUCCESS == code) {
code = tjsonAddIntegerToObject(pJson, jkLogicPlanResultDataOrder, pNode->resultDataOrder);
}
if (TSDB_CODE_SUCCESS == code) {
code = tjsonAddIntegerToObject(pJson, jkLogicPlanGroupAction, pNode->groupAction);
}
return code;
}
@ -541,6 +553,15 @@ static int32_t jsonToLogicPlanNode(const SJson* pJson, void* pObj) {
if (TSDB_CODE_SUCCESS == code) {
code = jsonToNodeObject(pJson, jkLogicPlanSlimit, &pNode->pSlimit);
}
if (TSDB_CODE_SUCCESS == code) {
tjsonGetNumberValue(pJson, jkLogicPlanRequireDataOrder, pNode->requireDataOrder, code);
}
if (TSDB_CODE_SUCCESS == code) {
tjsonGetNumberValue(pJson, jkLogicPlanResultDataOrder, pNode->resultDataOrder, code);
}
if (TSDB_CODE_SUCCESS == code) {
tjsonGetNumberValue(pJson, jkLogicPlanGroupAction, pNode->groupAction, code);
}
return code;
}

58
source/libs/parser/src/parInsert.c
View File

@ -434,8 +434,12 @@ static FORCE_INLINE int32_t checkAndTrimValue(SToken* pToken, char* tmpTokenBuf,
}
static bool isNullStr(SToken* pToken) {
return (pToken->type == TK_NULL) || ((pToken->type == TK_NK_STRING) && (pToken->n != 0) &&
(strncasecmp(TSDB_DATA_NULL_STR_L, pToken->z, pToken->n) == 0));
return ((pToken->type == TK_NK_STRING) && (pToken->n != 0) &&
(strncasecmp(TSDB_DATA_NULL_STR_L, pToken->z, pToken->n) == 0));
}
static bool isNullValue(int8_t dataType, SToken* pToken) {
return TK_NULL == pToken->type || (!IS_STR_DATA_TYPE(dataType) && isNullStr(pToken));
}
static FORCE_INLINE int32_t toDouble(SToken* pToken, double* value, char** endPtr) {
@ -461,7 +465,7 @@ static int32_t parseValueToken(char** end, SToken* pToken, SSchema* pSchema, int
return code;
}
if (isNullStr(pToken)) {
if (isNullValue(pSchema->type, pToken)) {
if (TSDB_DATA_TYPE_TIMESTAMP == pSchema->type && PRIMARYKEY_TIMESTAMP_COL_ID == pSchema->colId) {
return buildSyntaxErrMsg(pMsgBuf, "primary timestamp should not be null", pToken->z);
}
@ -735,11 +739,12 @@ static int32_t parseBoundColumns(SInsertParseContext* pCxt, SParsedDataColInfo*
return TSDB_CODE_SUCCESS;
}
static void buildCreateTbReq(SVCreateTbReq* pTbReq, const char* tname, STag* pTag, int64_t suid, const char* sname, SArray* tagName) {
static void buildCreateTbReq(SVCreateTbReq* pTbReq, const char* tname, STag* pTag, int64_t suid, const char* sname,
SArray* tagName) {
pTbReq->type = TD_CHILD_TABLE;
pTbReq->name = strdup(tname);
pTbReq->ctb.suid = suid;
if(sname) pTbReq->ctb.name = strdup(sname);
if (sname) pTbReq->ctb.name = strdup(sname);
pTbReq->ctb.pTag = (uint8_t*)pTag;
pTbReq->ctb.tagName = taosArrayDup(tagName);
pTbReq->commentLen = -1;
@ -753,7 +758,7 @@ static int32_t parseTagToken(char** end, SToken* pToken, SSchema* pSchema, int16
uint64_t uv;
char* endptr = NULL;
if (isNullStr(pToken)) {
if (isNullValue(pSchema->type, pToken)) {
if (TSDB_DATA_TYPE_TIMESTAMP == pSchema->type && PRIMARYKEY_TIMESTAMP_COL_ID == pSchema->colId) {
return buildSyntaxErrMsg(pMsgBuf, "primary timestamp should not be null", pToken->z);
}
@ -761,7 +766,7 @@ static int32_t parseTagToken(char** end, SToken* pToken, SSchema* pSchema, int16
return TSDB_CODE_SUCCESS;
}
// strcpy(val->colName, pSchema->name);
// strcpy(val->colName, pSchema->name);
val->cid = pSchema->colId;
val->type = pSchema->type;
@ -971,7 +976,7 @@ static int32_t parseTagsClause(SInsertParseContext* pCxt, SSchema* pSchema, uint
goto end;
}
if (!isNullStr(&sToken)) {
if (!isNullValue(pTagSchema->type, &sToken)) {
taosArrayPush(tagName, pTagSchema->name);
}
if (pTagSchema->type == TSDB_DATA_TYPE_JSON) {
@ -980,7 +985,7 @@ static int32_t parseTagsClause(SInsertParseContext* pCxt, SSchema* pSchema, uint
taosMemoryFree(tmpTokenBuf);
goto end;
}
if (isNullStr(&sToken)) {
if (isNullValue(pTagSchema->type, &sToken)) {
code = tTagNew(pTagVals, 1, true, &pTag);
} else {
code = parseJsontoTagData(sToken.z, pTagVals, &pTag, &pCxt->msg);
@ -1018,7 +1023,7 @@ static int32_t parseTagsClause(SInsertParseContext* pCxt, SSchema* pSchema, uint
end:
for (int i = 0; i < taosArrayGetSize(pTagVals); ++i) {
STagVal* p = (STagVal*)taosArrayGet(pTagVals, i);
if (IS_VAR_DATA_TYPE(p->type)) {
if (p->type == TSDB_DATA_TYPE_NCHAR) {
taosMemoryFree(p->pData);
}
}
@ -1200,7 +1205,7 @@ static int parseOneRow(SInsertParseContext* pCxt, STableDataBlocks* pDataBlocks,
*gotRow = true;
#ifdef TD_DEBUG_PRINT_ROW
STSchema* pSTSchema = tdGetSTSChemaFromSSChema(&schema, spd->numOfCols);
STSchema* pSTSchema = tdGetSTSChemaFromSSChema(schema, spd->numOfCols, 1);
tdSRowPrint(row, pSTSchema, __func__);
taosMemoryFree(pSTSchema);
#endif
@ -1321,7 +1326,11 @@ static int32_t parseCsvFile(SInsertParseContext* pCxt, TdFilePtr fp, STableDataB
static int32_t parseDataFromFile(SInsertParseContext* pCxt, SToken filePath, STableDataBlocks* dataBuf) {
char filePathStr[TSDB_FILENAME_LEN] = {0};
strncpy(filePathStr, filePath.z, filePath.n);
if (TK_NK_STRING == filePath.type) {
trimString(filePath.z, filePath.n, filePathStr, sizeof(filePathStr));
} else {
strncpy(filePathStr, filePath.z, filePath.n);
}
TdFilePtr fp = taosOpenFile(filePathStr, TD_FILE_READ | TD_FILE_STREAM);
if (NULL == fp) {
return TAOS_SYSTEM_ERROR(errno);
@ -1497,7 +1506,6 @@ static int32_t parseInsertBody(SInsertParseContext* pCxt) {
memset(&pCxt->tags, 0, sizeof(pCxt->tags));
pCxt->pVgroupsHashObj = NULL;
pCxt->pTableBlockHashObj = NULL;
pCxt->pTableMeta = NULL;
return TSDB_CODE_SUCCESS;
}
@ -1554,7 +1562,10 @@ int32_t parseInsertSql(SParseContext* pContext, SQuery** pQuery, SParseMetaCache
if (NULL == *pQuery) {
return TSDB_CODE_OUT_OF_MEMORY;
}
} else {
nodesDestroyNode((*pQuery)->pRoot);
}
(*pQuery)->execMode = QUERY_EXEC_MODE_SCHEDULE;
(*pQuery)->haveResultSet = false;
(*pQuery)->msgType = TDMT_VND_SUBMIT;
@ -1802,8 +1813,8 @@ int32_t qBuildStmtOutput(SQuery* pQuery, SHashObj* pVgHash, SHashObj* pBlockHash
return TSDB_CODE_SUCCESS;
}
int32_t qBindStmtTagsValue(void* pBlock, void* boundTags, int64_t suid, const char* sTableName, char* tName, TAOS_MULTI_BIND* bind,
char* msgBuf, int32_t msgBufLen) {
int32_t qBindStmtTagsValue(void* pBlock, void* boundTags, int64_t suid, const char* sTableName, char* tName,
TAOS_MULTI_BIND* bind, char* msgBuf, int32_t msgBufLen) {
STableDataBlocks* pDataBlock = (STableDataBlocks*)pBlock;
SMsgBuf pBuf = {.buf = msgBuf, .len = msgBufLen};
SParsedDataColInfo* tags = (SParsedDataColInfo*)boundTags;
@ -1854,7 +1865,7 @@ int32_t qBindStmtTagsValue(void* pBlock, void* boundTags, int64_t suid, const ch
}
} else {
STagVal val = {.cid = pTagSchema->colId, .type = pTagSchema->type};
// strcpy(val.colName, pTagSchema->name);
// strcpy(val.colName, pTagSchema->name);
if (pTagSchema->type == TSDB_DATA_TYPE_BINARY) {
val.pData = (uint8_t*)bind[c].buffer;
val.nData = colLen;
@ -1970,7 +1981,7 @@ int32_t qBindStmtColsValue(void* pBlock, TAOS_MULTI_BIND* bind, char* msgBuf, in
}
}
#ifdef TD_DEBUG_PRINT_ROW
STSchema* pSTSchema = tdGetSTSChemaFromSSChema(&pSchema, spd->numOfCols);
STSchema* pSTSchema = tdGetSTSChemaFromSSChema(pSchema, spd->numOfCols, 1);
tdSRowPrint(row, pSTSchema, __func__);
taosMemoryFree(pSTSchema);
#endif
@ -2055,7 +2066,7 @@ int32_t qBindStmtSingleColValue(void* pBlock, TAOS_MULTI_BIND* bind, char* msgBu
#ifdef TD_DEBUG_PRINT_ROW
if (rowEnd) {
STSchema* pSTSchema = tdGetSTSChemaFromSSChema(&pSchema, spd->numOfCols);
STSchema* pSTSchema = tdGetSTSChemaFromSSChema(pSchema, spd->numOfCols, 1);
tdSRowPrint(row, pSTSchema, __func__);
taosMemoryFree(pSTSchema);
}
@ -2245,7 +2256,8 @@ static int32_t smlBoundColumnData(SArray* cols, SParsedDataColInfo* pColList, SS
* @param msg
* @return int32_t
*/
static int32_t smlBuildTagRow(SArray* cols, SParsedDataColInfo* tags, SSchema* pSchema, STag** ppTag, SArray** tagName, SMsgBuf* msg) {
static int32_t smlBuildTagRow(SArray* cols, SParsedDataColInfo* tags, SSchema* pSchema, STag** ppTag, SArray** tagName,
SMsgBuf* msg) {
SArray* pTagArray = taosArrayInit(tags->numOfBound, sizeof(STagVal));
if (!pTagArray) {
return TSDB_CODE_TSC_OUT_OF_MEMORY;
@ -2262,7 +2274,7 @@ static int32_t smlBuildTagRow(SArray* cols, SParsedDataColInfo* tags, SSchema* p
taosArrayPush(*tagName, pTagSchema->name);
STagVal val = {.cid = pTagSchema->colId, .type = pTagSchema->type};
// strcpy(val.colName, pTagSchema->name);
// strcpy(val.colName, pTagSchema->name);
if (pTagSchema->type == TSDB_DATA_TYPE_BINARY) {
val.pData = (uint8_t*)kv->value;
val.nData = kv->length;
@ -2318,7 +2330,7 @@ int32_t smlBindData(void* handle, SArray* tags, SArray* colsSchema, SArray* cols
buildInvalidOperationMsg(&pBuf, "bound tags error");
return ret;
}
STag* pTag = NULL;
STag* pTag = NULL;
SArray* tagName = NULL;
ret = smlBuildTagRow(tags, &smlHandle->tableExecHandle.tags, pTagsSchema, &pTag, &tagName, &pBuf);
if (ret != TSDB_CODE_SUCCESS) {
@ -2402,9 +2414,9 @@ int32_t smlBindData(void* handle, SArray* tags, SArray* colsSchema, SArray* cols
} else {
int32_t colLen = kv->length;
if (pColSchema->type == TSDB_DATA_TYPE_TIMESTAMP) {
// uError("SML:data before:%ld, precision:%d", kv->i, pTableMeta->tableInfo.precision);
// uError("SML:data before:%ld, precision:%d", kv->i, pTableMeta->tableInfo.precision);
kv->i = convertTimePrecision(kv->i, TSDB_TIME_PRECISION_NANO, pTableMeta->tableInfo.precision);
// uError("SML:data after:%ld, precision:%d", kv->i, pTableMeta->tableInfo.precision);
// uError("SML:data after:%ld, precision:%d", kv->i, pTableMeta->tableInfo.precision);
}
if (IS_VAR_DATA_TYPE(kv->type)) {

233
source/libs/parser/src/parInsertData.c
View File

@ -19,6 +19,7 @@
#include "parInt.h"
#include "parUtil.h"
#include "querynodes.h"
#include "tRealloc.h"
#define IS_RAW_PAYLOAD(t) \
(((int)(t)) == PAYLOAD_TYPE_RAW) // 0: K-V payload for non-prepare insert, 1: rawPayload for prepare insert
@ -34,6 +35,32 @@ typedef struct SBlockKeyInfo {
SBlockKeyTuple* pKeyTuple;
} SBlockKeyInfo;
typedef struct {
int32_t index;
SArray* rowArray; // array of merged rows(mem allocated by tRealloc/free by tFree)
STSchema* pSchema;
int64_t tbUid; // suid for child table, uid for normal table
} SBlockRowMerger;
static FORCE_INLINE void tdResetSBlockRowMerger(SBlockRowMerger* pMerger) {
if (pMerger) {
pMerger->index = -1;
}
}
static void tdFreeSBlockRowMerger(SBlockRowMerger* pMerger) {
if (pMerger) {
int32_t size = taosArrayGetSize(pMerger->rowArray);
for (int32_t i = 0; i < size; ++i) {
tFree(*(void**)taosArrayGet(pMerger->rowArray, i));
}
taosArrayDestroy(pMerger->rowArray);
taosMemoryFreeClear(pMerger->pSchema);
taosMemoryFree(pMerger);
}
}
static int32_t rowDataCompar(const void* lhs, const void* rhs) {
TSKEY left = *(TSKEY*)lhs;
TSKEY right = *(TSKEY*)rhs;
@ -328,7 +355,7 @@ void sortRemoveDataBlockDupRowsRaw(STableDataBlocks* dataBuf) {
}
// data block is disordered, sort it in ascending order
int sortRemoveDataBlockDupRows(STableDataBlocks* dataBuf, SBlockKeyInfo* pBlkKeyInfo) {
static int sortRemoveDataBlockDupRows(STableDataBlocks* dataBuf, SBlockKeyInfo* pBlkKeyInfo) {
SSubmitBlk* pBlocks = (SSubmitBlk*)dataBuf->pData;
int16_t nRows = pBlocks->numOfRows;
@ -396,6 +423,201 @@ int sortRemoveDataBlockDupRows(STableDataBlocks* dataBuf, SBlockKeyInfo* pBlkKey
return 0;
}
static void* tdGetCurRowFromBlockMerger(SBlockRowMerger* pBlkRowMerger) {
if (pBlkRowMerger && (pBlkRowMerger->index >= 0)) {
ASSERT(pBlkRowMerger->index < taosArrayGetSize(pBlkRowMerger->rowArray));
return *(void**)taosArrayGet(pBlkRowMerger->rowArray, pBlkRowMerger->index);
}
return NULL;
}
static int32_t tdBlockRowMerge(STableMeta* pTableMeta, SBlockKeyTuple* pEndKeyTp, int32_t nDupRows,
SBlockRowMerger** pBlkRowMerger, int32_t rowSize) {
ASSERT(nDupRows > 1);
SBlockKeyTuple* pStartKeyTp = pEndKeyTp - (nDupRows - 1);
ASSERT(pStartKeyTp->skey == pEndKeyTp->skey);
// TODO: optimization if end row is all normal
#if 0
STSRow* pEndRow = (STSRow*)pEndKeyTp->payloadAddr;
if(isNormal(pEndRow)) { // set the end row if it is normal and return directly
pStartKeyTp->payloadAddr = pEndKeyTp->payloadAddr;
return TSDB_CODE_SUCCESS;
}
#endif
if (!(*pBlkRowMerger)) {
(*pBlkRowMerger) = taosMemoryCalloc(1, sizeof(**pBlkRowMerger));
if (!(*pBlkRowMerger)) {
terrno = TSDB_CODE_OUT_OF_MEMORY;
return TSDB_CODE_FAILED;
}
(*pBlkRowMerger)->index = -1;
if (!(*pBlkRowMerger)->rowArray) {
(*pBlkRowMerger)->rowArray = taosArrayInit(1, sizeof(void*));
if (!(*pBlkRowMerger)->rowArray) {
terrno = TSDB_CODE_OUT_OF_MEMORY;
return TSDB_CODE_FAILED;
}
}
}
if ((*pBlkRowMerger)->pSchema) {
if ((*pBlkRowMerger)->pSchema->version != pTableMeta->sversion) {
taosMemoryFreeClear((*pBlkRowMerger)->pSchema);
} else {
if ((*pBlkRowMerger)->tbUid != (pTableMeta->suid > 0 ? pTableMeta->suid : pTableMeta->uid)) {
taosMemoryFreeClear((*pBlkRowMerger)->pSchema);
}
}
}
if (!(*pBlkRowMerger)->pSchema) {
(*pBlkRowMerger)->pSchema =
tdGetSTSChemaFromSSChema(pTableMeta->schema, pTableMeta->tableInfo.numOfColumns, pTableMeta->sversion);
if (!(*pBlkRowMerger)->pSchema) {
terrno = TSDB_CODE_OUT_OF_MEMORY;
return TSDB_CODE_FAILED;
}
(*pBlkRowMerger)->tbUid = pTableMeta->suid > 0 ? pTableMeta->suid : pTableMeta->uid;
}
void* pDestRow = NULL;
++((*pBlkRowMerger)->index);
if ((*pBlkRowMerger)->index < taosArrayGetSize((*pBlkRowMerger)->rowArray)) {
void* pAlloc = *(void**)taosArrayGet((*pBlkRowMerger)->rowArray, (*pBlkRowMerger)->index);
if (tRealloc((uint8_t**)&pAlloc, rowSize) != 0) {
return TSDB_CODE_FAILED;
}
pDestRow = pAlloc;
} else {
if (tRealloc((uint8_t**)&pDestRow, rowSize) != 0) {
return TSDB_CODE_FAILED;
}
taosArrayPush((*pBlkRowMerger)->rowArray, &pDestRow);
}
// merge rows to pDestRow
STSchema* pSchema = (*pBlkRowMerger)->pSchema;
SArray* pArray = taosArrayInit(pSchema->numOfCols, sizeof(SColVal));
for (int32_t i = 0; i < pSchema->numOfCols; ++i) {
SColVal colVal = {0};
for (int32_t j = 0; j < nDupRows; ++j) {
tTSRowGetVal((pEndKeyTp - j)->payloadAddr, pSchema, i, &colVal);
if (!colVal.isNone) {
break;
}
}
taosArrayPush(pArray, &colVal);
}
if (tdSTSRowNew(pArray, pSchema, (STSRow**)&pDestRow) < 0) {
taosArrayDestroy(pArray);
return TSDB_CODE_FAILED;
}
taosArrayDestroy(pArray);
return TSDB_CODE_SUCCESS;
}
// data block is disordered, sort it in ascending order, and merge dup rows if exists
static int sortMergeDataBlockDupRows(STableDataBlocks* dataBuf, SBlockKeyInfo* pBlkKeyInfo,
SBlockRowMerger** ppBlkRowMerger) {
SSubmitBlk* pBlocks = (SSubmitBlk*)dataBuf->pData;
STableMeta* pTableMeta = dataBuf->pTableMeta;
int16_t nRows = pBlocks->numOfRows;
// size is less than the total size, since duplicated rows may be removed.
// allocate memory
size_t nAlloc = nRows * sizeof(SBlockKeyTuple);
if (pBlkKeyInfo->pKeyTuple == NULL || pBlkKeyInfo->maxBytesAlloc < nAlloc) {
char* tmp = taosMemoryRealloc(pBlkKeyInfo->pKeyTuple, nAlloc);
if (tmp == NULL) {
return TSDB_CODE_TSC_OUT_OF_MEMORY;
}
pBlkKeyInfo->pKeyTuple = (SBlockKeyTuple*)tmp;
pBlkKeyInfo->maxBytesAlloc = (int32_t)nAlloc;
}
memset(pBlkKeyInfo->pKeyTuple, 0, nAlloc);
tdResetSBlockRowMerger(*ppBlkRowMerger);
int32_t extendedRowSize = getExtendedRowSize(dataBuf);
SBlockKeyTuple* pBlkKeyTuple = pBlkKeyInfo->pKeyTuple;
char* pBlockData = pBlocks->data + pBlocks->schemaLen;
int n = 0;
while (n < nRows) {
pBlkKeyTuple->skey = TD_ROW_KEY((STSRow*)pBlockData);
pBlkKeyTuple->payloadAddr = pBlockData;
pBlkKeyTuple->index = n;
// next loop
pBlockData += extendedRowSize;
++pBlkKeyTuple;
++n;
}
if (!dataBuf->ordered) {
pBlkKeyTuple = pBlkKeyInfo->pKeyTuple;
taosSort(pBlkKeyTuple, nRows, sizeof(SBlockKeyTuple), rowDataComparStable);
pBlkKeyTuple = pBlkKeyInfo->pKeyTuple;
bool hasDup = false;
int32_t nextPos = 0;
int32_t i = 0;
int32_t j = 1;
while (j < nRows) {
TSKEY ti = (pBlkKeyTuple + i)->skey;
TSKEY tj = (pBlkKeyTuple + j)->skey;
if (ti == tj) {
++j;
continue;
}
if ((j - i) > 1) {
if (tdBlockRowMerge(pTableMeta, (pBlkKeyTuple + j - 1), j - i, ppBlkRowMerger, extendedRowSize) < 0) {
return TSDB_CODE_FAILED;
}
(pBlkKeyTuple + nextPos)->payloadAddr = tdGetCurRowFromBlockMerger(*ppBlkRowMerger);
if (!hasDup) {
hasDup = true;
}
i = j;
} else {
if (hasDup) {
memmove(pBlkKeyTuple + nextPos, pBlkKeyTuple + i, sizeof(SBlockKeyTuple));
}
++i;
}
++nextPos;
++j;
}
if ((j - i) > 1) {
ASSERT((pBlkKeyTuple + i)->skey == (pBlkKeyTuple + j - 1)->skey);
if (tdBlockRowMerge(pTableMeta, (pBlkKeyTuple + j - 1), j - i, ppBlkRowMerger, extendedRowSize) < 0) {
return TSDB_CODE_FAILED;
}
(pBlkKeyTuple + nextPos)->payloadAddr = tdGetCurRowFromBlockMerger(*ppBlkRowMerger);
} else if (hasDup) {
memmove(pBlkKeyTuple + nextPos, pBlkKeyTuple + i, sizeof(SBlockKeyTuple));
}
dataBuf->ordered = true;
pBlocks->numOfRows = nextPos + 1;
}
dataBuf->size = sizeof(SSubmitBlk) + pBlocks->numOfRows * extendedRowSize;
dataBuf->prevTS = INT64_MIN;
return TSDB_CODE_SUCCESS;
}
// Erase the empty space reserved for binary data
static int trimDataBlock(void* pDataBlock, STableDataBlocks* pTableDataBlock, SBlockKeyTuple* blkKeyTuple,
bool isRawPayload) {
@ -464,6 +686,8 @@ int32_t mergeTableDataBlocks(SHashObj* pHashObj, uint8_t payloadType, SArray** p
STableDataBlocks** p = taosHashIterate(pHashObj, NULL);
STableDataBlocks* pOneTableBlock = *p;
SBlockKeyInfo blkKeyInfo = {0}; // share by pOneTableBlock
SBlockRowMerger *pBlkRowMerger = NULL;
while (pOneTableBlock) {
SSubmitBlk* pBlocks = (SSubmitBlk*)pOneTableBlock->pData;
if (pBlocks->numOfRows > 0) {
@ -473,6 +697,7 @@ int32_t mergeTableDataBlocks(SHashObj* pHashObj, uint8_t payloadType, SArray** p
getDataBlockFromList(pVnodeDataBlockHashList, &pOneTableBlock->vgId, sizeof(pOneTableBlock->vgId), TSDB_PAYLOAD_SIZE, INSERT_HEAD_SIZE, 0,
pOneTableBlock->pTableMeta, &dataBuf, pVnodeDataBlockList, NULL);
if (ret != TSDB_CODE_SUCCESS) {
tdFreeSBlockRowMerger(pBlkRowMerger);
taosHashCleanup(pVnodeDataBlockHashList);
destroyBlockArrayList(pVnodeDataBlockList);
taosMemoryFreeClear(blkKeyInfo.pKeyTuple);
@ -490,6 +715,7 @@ int32_t mergeTableDataBlocks(SHashObj* pHashObj, uint8_t payloadType, SArray** p
if (tmp != NULL) {
dataBuf->pData = tmp;
} else { // failed to allocate memory, free already allocated memory and return error code
tdFreeSBlockRowMerger(pBlkRowMerger);
taosHashCleanup(pVnodeDataBlockHashList);
destroyBlockArrayList(pVnodeDataBlockList);
taosMemoryFreeClear(dataBuf->pData);
@ -501,7 +727,8 @@ int32_t mergeTableDataBlocks(SHashObj* pHashObj, uint8_t payloadType, SArray** p
if (isRawPayload) {
sortRemoveDataBlockDupRowsRaw(pOneTableBlock);
} else {
if ((code = sortRemoveDataBlockDupRows(pOneTableBlock, &blkKeyInfo)) != 0) {
if ((code = sortMergeDataBlockDupRows(pOneTableBlock, &blkKeyInfo, &pBlkRowMerger)) != 0) {
tdFreeSBlockRowMerger(pBlkRowMerger);
taosHashCleanup(pVnodeDataBlockHashList);
destroyBlockArrayList(pVnodeDataBlockList);
taosMemoryFreeClear(dataBuf->pData);
@ -529,6 +756,7 @@ int32_t mergeTableDataBlocks(SHashObj* pHashObj, uint8_t payloadType, SArray** p
}
// free the table data blocks;
tdFreeSBlockRowMerger(pBlkRowMerger);
taosHashCleanup(pVnodeDataBlockHashList);
taosMemoryFreeClear(blkKeyInfo.pKeyTuple);
*pVgDataBlocks = pVnodeDataBlockList;
@ -678,6 +906,7 @@ void qFreeStmtDataBlock(void* pDataBlock) {
return;
}
taosMemoryFreeClear(((STableDataBlocks*)pDataBlock)->pTableMeta);
taosMemoryFreeClear(((STableDataBlocks*)pDataBlock)->pData);
taosMemoryFreeClear(pDataBlock);
}

2
source/libs/parser/src/parTranslater.c
View File

@ -1089,7 +1089,7 @@ static int32_t translateScanPseudoColumnFunc(STranslateContext* pCxt, SFunctionN
return TSDB_CODE_SUCCESS;
}
if (0 == LIST_LENGTH(pFunc->pParameterList)) {
if (!isSelectStmt(pCxt->pCurrStmt) ||
if (!isSelectStmt(pCxt->pCurrStmt) || NULL == ((SSelectStmt*)pCxt->pCurrStmt)->pFromTable ||
QUERY_NODE_REAL_TABLE != nodeType(((SSelectStmt*)pCxt->pCurrStmt)->pFromTable)) {
return generateSyntaxErrMsg(&pCxt->msgBuf, TSDB_CODE_PAR_INVALID_TBNAME);
}

6
source/libs/parser/src/parUtil.c
View File

@ -866,13 +866,15 @@ STableCfg* tableCfgDup(STableCfg* pCfg) {
memcpy(pNew, pCfg, sizeof(*pNew));
if (NULL != pNew->pComment) {
pNew->pComment = strdup(pNew->pComment);
pNew->pComment = taosMemoryCalloc(pNew->commentLen + 1, 1);
memcpy(pNew->pComment, pCfg->pComment, pNew->commentLen);
}
if (NULL != pNew->pFuncs) {
pNew->pFuncs = taosArrayDup(pNew->pFuncs);
}
if (NULL != pNew->pTags) {
pNew->pTags = strdup(pNew->pTags);
pNew->pTags = taosMemoryCalloc(pNew->tagsLen + 1, 1);
memcpy(pNew->pTags, pCfg->pTags, pNew->tagsLen);
}
int32_t schemaSize = (pCfg->numOfColumns + pCfg->numOfTags) * sizeof(SSchema);

7
source/libs/parser/src/parser.c
View File

@ -36,7 +36,7 @@ bool qIsInsertValuesSql(const char* pStr, size_t length) {
pStr += index;
index = 0;
t = tStrGetToken((char*)pStr, &index, false);
if (TK_USING == t.type || TK_VALUES == t.type) {
if (TK_USING == t.type || TK_VALUES == t.type || TK_FILE == t.type) {
return true;
} else if (TK_SELECT == t.type) {
return false;
@ -82,11 +82,16 @@ static int32_t parseSqlSyntax(SParseContext* pCxt, SQuery** pQuery, SParseMetaCa
}
static int32_t setValueByBindParam(SValueNode* pVal, TAOS_MULTI_BIND* pParam) {
if (IS_VAR_DATA_TYPE(pVal->node.resType.type)) {
taosMemoryFreeClear(pVal->datum.p);
}
if (pParam->is_null && 1 == *(pParam->is_null)) {
pVal->node.resType.type = TSDB_DATA_TYPE_NULL;
pVal->node.resType.bytes = tDataTypes[TSDB_DATA_TYPE_NULL].bytes;
return TSDB_CODE_SUCCESS;
}
int32_t inputSize = (NULL != pParam->length ? *(pParam->length) : tDataTypes[pParam->buffer_type].bytes);
pVal->node.resType.type = pParam->buffer_type;
pVal->node.resType.bytes = inputSize;

7
source/libs/parser/test/parSelectTest.cpp
View File

@ -444,4 +444,11 @@ TEST_F(ParserSelectTest, withoutFrom) {
run("SELECT USER()");
}
TEST_F(ParserSelectTest, withoutFromSemanticCheck) {
useDb("root", "test");
run("SELECT c1", TSDB_CODE_PAR_INVALID_COLUMN);
run("SELECT TBNAME", TSDB_CODE_PAR_INVALID_TBNAME);
}
} // namespace ParserTest

1
source/libs/planner/inc/planInt.h
View File

@ -35,6 +35,7 @@ int32_t generateUsageErrMsg(char* pBuf, int32_t len, int32_t errCode, ...);
int32_t createColumnByRewriteExprs(SNodeList* pExprs, SNodeList** pList);
int32_t createColumnByRewriteExpr(SNode* pExpr, SNodeList** pList);
int32_t replaceLogicNode(SLogicSubplan* pSubplan, SLogicNode* pOld, SLogicNode* pNew);
int32_t adjustLogicNodeDataRequirement(SLogicNode* pNode, EDataOrderLevel requirement);
int32_t createLogicPlan(SPlanContext* pCxt, SLogicSubplan** pLogicSubplan);
int32_t optimizeLogicPlan(SPlanContext* pCxt, SLogicSubplan* pLogicSubplan);

62
source/libs/planner/src/planLogicCreater.c
View File

@ -250,6 +250,9 @@ static int32_t createScanLogicNode(SLogicPlanContext* pCxt, SSelectStmt* pSelect
SScanLogicNode* pScan = NULL;
int32_t code = makeScanLogicNode(pCxt, pRealTable, pSelect->hasRepeatScanFuncs, (SLogicNode**)&pScan);
pScan->node.groupAction = GROUP_ACTION_NONE;
pScan->node.resultDataOrder = DATA_ORDER_LEVEL_IN_BLOCK;
// set columns to scan
if (TSDB_CODE_SUCCESS == code) {
code = nodesCollectColumns(pSelect, SQL_CLAUSE_FROM, pRealTable->table.tableAlias, COLLECT_COL_TYPE_COL,
@ -336,6 +339,9 @@ static int32_t createJoinLogicNode(SLogicPlanContext* pCxt, SSelectStmt* pSelect
pJoin->joinType = pJoinTable->joinType;
pJoin->isSingleTableJoin = pJoinTable->table.singleTable;
pJoin->node.groupAction = GROUP_ACTION_CLEAR;
pJoin->node.requireDataOrder = DATA_ORDER_LEVEL_GLOBAL;
pJoin->node.requireDataOrder = DATA_ORDER_LEVEL_GLOBAL;
int32_t code = TSDB_CODE_SUCCESS;
@ -472,6 +478,9 @@ static int32_t createAggLogicNode(SLogicPlanContext* pCxt, SSelectStmt* pSelect,
}
pAgg->hasLastRow = pSelect->hasLastRowFunc;
pAgg->node.groupAction = GROUP_ACTION_SET;
pAgg->node.requireDataOrder = DATA_ORDER_LEVEL_NONE;
pAgg->node.resultDataOrder = DATA_ORDER_LEVEL_NONE;
int32_t code = TSDB_CODE_SUCCESS;
@ -540,6 +549,10 @@ static int32_t createIndefRowsFuncLogicNode(SLogicPlanContext* pCxt, SSelectStmt
pIdfRowsFunc->isTailFunc = pSelect->hasTailFunc;
pIdfRowsFunc->isUniqueFunc = pSelect->hasUniqueFunc;
pIdfRowsFunc->isTimeLineFunc = pSelect->hasTimeLineFunc;
pIdfRowsFunc->node.groupAction = GROUP_ACTION_KEEP;
pIdfRowsFunc->node.requireDataOrder =
pIdfRowsFunc->isTimeLineFunc ? DATA_ORDER_LEVEL_IN_GROUP : DATA_ORDER_LEVEL_NONE;
pIdfRowsFunc->node.resultDataOrder = pIdfRowsFunc->node.requireDataOrder;
// indefinite rows functions and _select_values functions
int32_t code = nodesCollectFuncs(pSelect, SQL_CLAUSE_SELECT, fmIsVectorFunc, &pIdfRowsFunc->pFuncs);
@ -571,6 +584,10 @@ static int32_t createInterpFuncLogicNode(SLogicPlanContext* pCxt, SSelectStmt* p
return TSDB_CODE_OUT_OF_MEMORY;
}
pInterpFunc->node.groupAction = GROUP_ACTION_KEEP;
pInterpFunc->node.requireDataOrder = DATA_ORDER_LEVEL_IN_GROUP;
pInterpFunc->node.resultDataOrder = pInterpFunc->node.requireDataOrder;
int32_t code = nodesCollectFuncs(pSelect, SQL_CLAUSE_SELECT, fmIsInterpFunc, &pInterpFunc->pFuncs);
if (TSDB_CODE_SUCCESS == code) {
code = rewriteExprsForSelect(pInterpFunc->pFuncs, pSelect, SQL_CLAUSE_SELECT);
@ -642,10 +659,12 @@ static int32_t createWindowLogicNodeByState(SLogicPlanContext* pCxt, SStateWindo
}
pWindow->winType = WINDOW_TYPE_STATE;
pWindow->node.groupAction = GROUP_ACTION_KEEP;
pWindow->node.requireDataOrder = pCxt->pPlanCxt->streamQuery ? DATA_ORDER_LEVEL_IN_BLOCK : DATA_ORDER_LEVEL_IN_GROUP;
pWindow->node.resultDataOrder = DATA_ORDER_LEVEL_IN_GROUP;
pWindow->pStateExpr = nodesCloneNode(pState->pExpr);
pWindow->pTspk = nodesCloneNode(pState->pCol);
if (NULL == pWindow->pTspk) {
if (NULL == pWindow->pStateExpr || NULL == pWindow->pTspk) {
nodesDestroyNode((SNode*)pWindow);
return TSDB_CODE_OUT_OF_MEMORY;
}
@ -663,6 +682,9 @@ static int32_t createWindowLogicNodeBySession(SLogicPlanContext* pCxt, SSessionW
pWindow->winType = WINDOW_TYPE_SESSION;
pWindow->sessionGap = ((SValueNode*)pSession->pGap)->datum.i;
pWindow->windowAlgo = pCxt->pPlanCxt->streamQuery ? SESSION_ALGO_STREAM_SINGLE : SESSION_ALGO_MERGE;
pWindow->node.groupAction = GROUP_ACTION_KEEP;
pWindow->node.requireDataOrder = pCxt->pPlanCxt->streamQuery ? DATA_ORDER_LEVEL_IN_BLOCK : DATA_ORDER_LEVEL_IN_GROUP;
pWindow->node.resultDataOrder = DATA_ORDER_LEVEL_IN_GROUP;
pWindow->pTspk = nodesCloneNode((SNode*)pSession->pCol);
if (NULL == pWindow->pTspk) {
@ -689,6 +711,9 @@ static int32_t createWindowLogicNodeByInterval(SLogicPlanContext* pCxt, SInterva
pWindow->slidingUnit =
(NULL != pInterval->pSliding ? ((SValueNode*)pInterval->pSliding)->unit : pWindow->intervalUnit);
pWindow->windowAlgo = pCxt->pPlanCxt->streamQuery ? INTERVAL_ALGO_STREAM_SINGLE : INTERVAL_ALGO_HASH;
pWindow->node.groupAction = GROUP_ACTION_KEEP;
pWindow->node.requireDataOrder = DATA_ORDER_LEVEL_IN_BLOCK;
pWindow->node.resultDataOrder = DATA_ORDER_LEVEL_IN_GROUP;
pWindow->pTspk = nodesCloneNode(pInterval->pCol);
if (NULL == pWindow->pTspk) {
@ -734,6 +759,10 @@ static int32_t createFillLogicNode(SLogicPlanContext* pCxt, SSelectStmt* pSelect
return TSDB_CODE_OUT_OF_MEMORY;
}
pFill->node.groupAction = GROUP_ACTION_KEEP;
pFill->node.requireDataOrder = DATA_ORDER_LEVEL_IN_GROUP;
pFill->node.resultDataOrder = DATA_ORDER_LEVEL_IN_GROUP;
int32_t code = nodesCollectColumns(pSelect, SQL_CLAUSE_WINDOW, NULL, COLLECT_COL_TYPE_ALL, &pFill->node.pTargets);
if (TSDB_CODE_SUCCESS == code && NULL == pFill->node.pTargets) {
code = nodesListMakeStrictAppend(&pFill->node.pTargets,
@ -768,6 +797,9 @@ static int32_t createSortLogicNode(SLogicPlanContext* pCxt, SSelectStmt* pSelect
}
pSort->groupSort = pSelect->groupSort;
pSort->node.groupAction = pSort->groupSort ? GROUP_ACTION_KEEP : GROUP_ACTION_CLEAR;
pSort->node.requireDataOrder = DATA_ORDER_LEVEL_NONE;
pSort->node.resultDataOrder = pSort->groupSort ? DATA_ORDER_LEVEL_IN_GROUP : DATA_ORDER_LEVEL_GLOBAL;
int32_t code = nodesCollectColumns(pSelect, SQL_CLAUSE_ORDER_BY, NULL, COLLECT_COL_TYPE_ALL, &pSort->node.pTargets);
if (TSDB_CODE_SUCCESS == code && NULL == pSort->node.pTargets) {
@ -818,6 +850,9 @@ static int32_t createProjectLogicNode(SLogicPlanContext* pCxt, SSelectStmt* pSel
TSWAP(pProject->node.pLimit, pSelect->pLimit);
TSWAP(pProject->node.pSlimit, pSelect->pSlimit);
pProject->node.groupAction = GROUP_ACTION_CLEAR;
pProject->node.requireDataOrder = DATA_ORDER_LEVEL_NONE;
pProject->node.resultDataOrder = DATA_ORDER_LEVEL_NONE;
int32_t code = TSDB_CODE_SUCCESS;
@ -850,6 +885,10 @@ static int32_t createPartitionLogicNode(SLogicPlanContext* pCxt, SSelectStmt* pS
return TSDB_CODE_OUT_OF_MEMORY;
}
pPartition->node.groupAction = GROUP_ACTION_SET;
pPartition->node.requireDataOrder = DATA_ORDER_LEVEL_NONE;
pPartition->node.resultDataOrder = DATA_ORDER_LEVEL_NONE;
int32_t code =
nodesCollectColumns(pSelect, SQL_CLAUSE_PARTITION_BY, NULL, COLLECT_COL_TYPE_ALL, &pPartition->node.pTargets);
if (TSDB_CODE_SUCCESS == code && NULL == pPartition->node.pTargets) {
@ -882,11 +921,23 @@ static int32_t createDistinctLogicNode(SLogicPlanContext* pCxt, SSelectStmt* pSe
return TSDB_CODE_OUT_OF_MEMORY;
}
pAgg->node.groupAction = GROUP_ACTION_SET;
pAgg->node.requireDataOrder = DATA_ORDER_LEVEL_NONE;
pAgg->node.resultDataOrder = DATA_ORDER_LEVEL_NONE;
int32_t code = TSDB_CODE_SUCCESS;
// set grouyp keys, agg funcs and having conditions
pAgg->pGroupKeys = nodesCloneList(pSelect->pProjectionList);
if (NULL == pAgg->pGroupKeys) {
code = TSDB_CODE_OUT_OF_MEMORY;
SNodeList* pGroupKeys = NULL;
SNode* pProjection = NULL;
FOREACH(pProjection, pSelect->pProjectionList) {
code = nodesListMakeStrictAppend(&pGroupKeys, createGroupingSetNode(pProjection));
if (TSDB_CODE_SUCCESS != code) {
nodesDestroyList(pGroupKeys);
break;
}
}
if (TSDB_CODE_SUCCESS == code) {
pAgg->pGroupKeys = pGroupKeys;
}
// rewrite the expression in subsequent clauses
@ -1361,6 +1412,7 @@ int32_t createLogicPlan(SPlanContext* pCxt, SLogicSubplan** pLogicSubplan) {
if (TSDB_CODE_SUCCESS == code) {
setLogicNodeParent(pSubplan->pNode);
setLogicSubplanType(cxt.hasScan, pSubplan);
code = adjustLogicNodeDataRequirement(pSubplan->pNode, DATA_ORDER_LEVEL_NONE);
}
if (TSDB_CODE_SUCCESS == code) {

94
source/libs/planner/src/planOptimizer.c
View File

@ -1579,6 +1579,34 @@ static bool eliminateProjOptMayBeOptimized(SLogicNode* pNode) {
return eliminateProjOptCheckProjColumnNames(pProjectNode);
}
typedef struct CheckNewChildTargetsCxt {
SNodeList* pNewChildTargets;
bool canUse;
} CheckNewChildTargetsCxt;
static EDealRes eliminateProjOptCanUseNewChildTargetsImpl(SNode* pNode, void* pContext) {
if (QUERY_NODE_COLUMN == nodeType(pNode)) {
CheckNewChildTargetsCxt* pCxt = pContext;
SNode* pTarget = NULL;
FOREACH(pTarget, pCxt->pNewChildTargets) {
if (!nodesEqualNode(pTarget, pNode)) {
pCxt->canUse = false;
return DEAL_RES_END;
}
}
}
return DEAL_RES_CONTINUE;
}
static bool eliminateProjOptCanUseNewChildTargets(SLogicNode* pChild, SNodeList* pNewChildTargets) {
if (NULL == pChild->pConditions) {
return true;
}
CheckNewChildTargetsCxt cxt = {.pNewChildTargets = pNewChildTargets, .canUse = true};
nodesWalkExpr(pChild->pConditions, eliminateProjOptCanUseNewChildTargetsImpl, &cxt);
return cxt.canUse;
}
static int32_t eliminateProjOptimizeImpl(SOptimizeContext* pCxt, SLogicSubplan* pLogicSubplan,
SProjectLogicNode* pProjectNode) {
SLogicNode* pChild = (SLogicNode*)nodesListGetNode(pProjectNode->node.pChildren, 0);
@ -1588,14 +1616,19 @@ static int32_t eliminateProjOptimizeImpl(SOptimizeContext* pCxt, SLogicSubplan*
FOREACH(pProjection, pProjectNode->pProjections) {
SNode* pChildTarget = NULL;
FOREACH(pChildTarget, pChild->pTargets) {
if (strcmp(((SColumnNode*)pProjection)->colName, ((SColumnNode*)pChildTarget)->colName) == 0) {
if (0 == strcmp(((SColumnNode*)pProjection)->colName, ((SColumnNode*)pChildTarget)->colName)) {
nodesListAppend(pNewChildTargets, nodesCloneNode(pChildTarget));
break;
}
}
}
nodesDestroyList(pChild->pTargets);
pChild->pTargets = pNewChildTargets;
if (eliminateProjOptCanUseNewChildTargets(pChild, pNewChildTargets)) {
nodesDestroyList(pChild->pTargets);
pChild->pTargets = pNewChildTargets;
} else {
nodesDestroyList(pNewChildTargets);
return TSDB_CODE_SUCCESS;
}
int32_t code = replaceLogicNode(pLogicSubplan, (SLogicNode*)pProjectNode, pChild);
if (TSDB_CODE_SUCCESS == code) {
@ -1873,6 +1906,8 @@ static int32_t rewriteUniqueOptCreateAgg(SIndefRowsFuncLogicNode* pIndef, SLogic
TSWAP(pAgg->node.pChildren, pIndef->node.pChildren);
optResetParent((SLogicNode*)pAgg);
pAgg->node.precision = pIndef->node.precision;
pAgg->node.requireDataOrder = DATA_ORDER_LEVEL_IN_BLOCK; // first function requirement
pAgg->node.resultDataOrder = DATA_ORDER_LEVEL_NONE;
int32_t code = TSDB_CODE_SUCCESS;
bool hasSelectPrimaryKey = false;
@ -1945,6 +1980,8 @@ static int32_t rewriteUniqueOptCreateProject(SIndefRowsFuncLogicNode* pIndef, SL
TSWAP(pProject->node.pTargets, pIndef->node.pTargets);
pProject->node.precision = pIndef->node.precision;
pProject->node.requireDataOrder = DATA_ORDER_LEVEL_NONE;
pProject->node.resultDataOrder = DATA_ORDER_LEVEL_NONE;
int32_t code = TSDB_CODE_SUCCESS;
SNode* pNode = NULL;
@ -1973,12 +2010,17 @@ static int32_t rewriteUniqueOptimizeImpl(SOptimizeContext* pCxt, SLogicSubplan*
}
if (TSDB_CODE_SUCCESS == code) {
code = nodesListMakeAppend(&pProject->pChildren, (SNode*)pAgg);
pAgg->pParent = pProject;
pAgg = NULL;
}
if (TSDB_CODE_SUCCESS == code) {
pAgg->pParent = pProject;
pAgg = NULL;
code = replaceLogicNode(pLogicSubplan, (SLogicNode*)pIndef, pProject);
}
if (TSDB_CODE_SUCCESS == code) {
code = adjustLogicNodeDataRequirement(
pProject, NULL == pProject->pParent ? DATA_ORDER_LEVEL_NONE : pProject->pParent->requireDataOrder);
pProject = NULL;
}
if (TSDB_CODE_SUCCESS == code) {
nodesDestroyNode((SNode*)pIndef);
} else {
@ -2145,11 +2187,20 @@ static bool tagScanMayBeOptimized(SLogicNode* pNode) {
}
SAggLogicNode* pAgg = (SAggLogicNode*)(pNode->pParent);
if (NULL == pAgg->pGroupKeys || NULL != pAgg->pAggFuncs ||
planOptNodeListHasCol(pAgg->pGroupKeys) || !planOptNodeListHasTbname(pAgg->pGroupKeys)) {
if (NULL == pAgg->pGroupKeys || NULL != pAgg->pAggFuncs || planOptNodeListHasCol(pAgg->pGroupKeys) ||
!planOptNodeListHasTbname(pAgg->pGroupKeys)) {
return false;
}
SNode* pGroupKey = NULL;
FOREACH(pGroupKey, pAgg->pGroupKeys) {
SNode* pGroup = NULL;
FOREACH(pGroup, ((SGroupingSetNode*)pGroupKey)->pParameterList) {
if (QUERY_NODE_COLUMN != nodeType(pGroup)) {
return false;
}
}
}
return true;
}
@ -2162,15 +2213,32 @@ static int32_t tagScanOptimize(SOptimizeContext* pCxt, SLogicSubplan* pLogicSubp
pScanNode->scanType = SCAN_TYPE_TAG;
SNode* pTarget = NULL;
FOREACH(pTarget, pScanNode->node.pTargets) {
if (PRIMARYKEY_TIMESTAMP_COL_ID == ((SColumnNode*)(pTarget))->colId) {
ERASE_NODE(pScanNode->node.pTargets);
break;
}
if (PRIMARYKEY_TIMESTAMP_COL_ID == ((SColumnNode*)(pTarget))->colId) {
ERASE_NODE(pScanNode->node.pTargets);
break;
}
}
NODES_DESTORY_LIST(pScanNode->pScanCols);
SLogicNode* pAgg = pScanNode->node.pParent;
if (NULL == pAgg->pParent) {
SNodeList* pScanTargets = nodesMakeList();
SNode* pAggTarget = NULL;
FOREACH(pAggTarget, pAgg->pTargets) {
SNode* pScanTarget = NULL;
FOREACH(pScanTarget, pScanNode->node.pTargets) {
if (0 == strcmp(((SColumnNode*)pAggTarget)->colName, ((SColumnNode*)pAggTarget)->colName)) {
nodesListAppend(pScanTargets, nodesCloneNode(pScanTarget));
break;
}
}
}
nodesDestroyList(pScanNode->node.pTargets);
pScanNode->node.pTargets = pScanTargets;
}
int32_t code = replaceLogicNode(pLogicSubplan, pAgg, (SLogicNode*)pScanNode);
if (TSDB_CODE_SUCCESS == code) {
NODES_CLEAR_LIST(pAgg->pChildren);

13
source/libs/planner/src/planPhysiCreater.c
View File

@ -974,6 +974,17 @@ static int32_t createInterpFuncPhysiNode(SPhysiPlanContext* pCxt, SNodeList* pCh
return code;
}
static bool projectCanMergeDataBlock(SProjectLogicNode* pProject) {
if (DATA_ORDER_LEVEL_NONE == pProject->node.resultDataOrder) {
return true;
}
if (1 != LIST_LENGTH(pProject->node.pChildren)) {
return false;
}
SLogicNode* pChild = (SLogicNode*)nodesListGetNode(pProject->node.pChildren, 0);
return DATA_ORDER_LEVEL_GLOBAL == pChild->resultDataOrder ? true : false;
}
static int32_t createProjectPhysiNode(SPhysiPlanContext* pCxt, SNodeList* pChildren,
SProjectLogicNode* pProjectLogicNode, SPhysiNode** pPhyNode) {
SProjectPhysiNode* pProject =
@ -982,6 +993,8 @@ static int32_t createProjectPhysiNode(SPhysiPlanContext* pCxt, SNodeList* pChild
return TSDB_CODE_OUT_OF_MEMORY;
}
pProject->mergeDataBlock = projectCanMergeDataBlock(pProjectLogicNode);
int32_t code = TSDB_CODE_SUCCESS;
if (0 == LIST_LENGTH(pChildren)) {
pProject->pProjections = nodesCloneList(pProjectLogicNode->pProjections);

3
source/libs/planner/src/planSpliter.c
View File

@ -657,6 +657,9 @@ static int32_t stbSplSplitWindowForPartTable(SSplitContext* pCxt, SStableSplitIn
return TSDB_CODE_SUCCESS;
}
if (NULL != pInfo->pSplitNode->pParent && QUERY_NODE_LOGIC_PLAN_FILL == nodeType(pInfo->pSplitNode->pParent)) {
pInfo->pSplitNode = pInfo->pSplitNode->pParent;
}
SExchangeLogicNode* pExchange = NULL;
int32_t code = splCreateExchangeNode(pCxt, pInfo->pSplitNode, &pExchange);
if (TSDB_CODE_SUCCESS == code) {

183
source/libs/planner/src/planUtil.c
View File

@ -13,6 +13,7 @@
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include "functionMgt.h"
#include "planInt.h"
static char* getUsageErrFormat(int32_t errCode) {
@ -121,3 +122,185 @@ int32_t replaceLogicNode(SLogicSubplan* pSubplan, SLogicNode* pOld, SLogicNode*
}
return TSDB_CODE_PLAN_INTERNAL_ERROR;
}
static int32_t adjustScanDataRequirement(SScanLogicNode* pScan, EDataOrderLevel requirement) {
if (SCAN_TYPE_TABLE != pScan->scanType || SCAN_TYPE_TABLE_MERGE != pScan->scanType) {
return TSDB_CODE_SUCCESS;
}
// The lowest sort level of scan output data is DATA_ORDER_LEVEL_IN_BLOCK
if (requirement < DATA_ORDER_LEVEL_IN_BLOCK) {
requirement = DATA_ORDER_LEVEL_IN_BLOCK;
}
if (DATA_ORDER_LEVEL_IN_BLOCK == requirement) {
pScan->scanType = SCAN_TYPE_TABLE;
} else if (TSDB_SUPER_TABLE == pScan->tableType) {
pScan->scanType = SCAN_TYPE_TABLE_MERGE;
}
pScan->node.resultDataOrder = requirement;
return TSDB_CODE_SUCCESS;
}
static int32_t adjustJoinDataRequirement(SJoinLogicNode* pJoin, EDataOrderLevel requirement) {
// The lowest sort level of join input and output data is DATA_ORDER_LEVEL_GLOBAL
return TSDB_CODE_SUCCESS;
}
static bool isKeepOrderAggFunc(SNodeList* pFuncs) {
SNode* pFunc = NULL;
FOREACH(pFunc, pFuncs) {
if (!fmIsKeepOrderFunc(((SFunctionNode*)pFunc)->funcId)) {
return false;
}
}
return true;
}
static int32_t adjustAggDataRequirement(SAggLogicNode* pAgg, EDataOrderLevel requirement) {
// The sort level of agg with group by output data can only be DATA_ORDER_LEVEL_NONE
if (requirement > DATA_ORDER_LEVEL_NONE && (NULL != pAgg->pGroupKeys || !isKeepOrderAggFunc(pAgg->pAggFuncs))) {
return TSDB_CODE_PLAN_INTERNAL_ERROR;
}
pAgg->node.resultDataOrder = requirement;
pAgg->node.requireDataOrder = requirement;
return TSDB_CODE_SUCCESS;
}
static int32_t adjustProjectDataRequirement(SProjectLogicNode* pProject, EDataOrderLevel requirement) {
pProject->node.resultDataOrder = requirement;
pProject->node.requireDataOrder = requirement;
return TSDB_CODE_SUCCESS;
}
static int32_t adjustIntervalDataRequirement(SWindowLogicNode* pWindow, EDataOrderLevel requirement) {
// The lowest sort level of interval output data is DATA_ORDER_LEVEL_IN_GROUP
if (requirement < DATA_ORDER_LEVEL_IN_GROUP) {
requirement = DATA_ORDER_LEVEL_IN_GROUP;
}
// The sort level of interval input data is always DATA_ORDER_LEVEL_IN_BLOCK
pWindow->node.resultDataOrder = requirement;
return TSDB_CODE_SUCCESS;
}
static int32_t adjustSessionDataRequirement(SWindowLogicNode* pWindow, EDataOrderLevel requirement) {
if (requirement <= pWindow->node.resultDataOrder) {
return TSDB_CODE_SUCCESS;
}
pWindow->node.resultDataOrder = requirement;
pWindow->node.requireDataOrder = requirement;
return TSDB_CODE_SUCCESS;
}
static int32_t adjustStateDataRequirement(SWindowLogicNode* pWindow, EDataOrderLevel requirement) {
if (requirement <= pWindow->node.resultDataOrder) {
return TSDB_CODE_SUCCESS;
}
pWindow->node.resultDataOrder = requirement;
pWindow->node.requireDataOrder = requirement;
return TSDB_CODE_SUCCESS;
}
static int32_t adjustWindowDataRequirement(SWindowLogicNode* pWindow, EDataOrderLevel requirement) {
switch (pWindow->winType) {
case WINDOW_TYPE_INTERVAL:
return adjustIntervalDataRequirement(pWindow, requirement);
case WINDOW_TYPE_SESSION:
return adjustSessionDataRequirement(pWindow, requirement);
case WINDOW_TYPE_STATE:
return adjustStateDataRequirement(pWindow, requirement);
default:
break;
}
return TSDB_CODE_PLAN_INTERNAL_ERROR;
}
static int32_t adjustFillDataRequirement(SFillLogicNode* pFill, EDataOrderLevel requirement) {
if (requirement <= pFill->node.requireDataOrder) {
return TSDB_CODE_SUCCESS;
}
pFill->node.resultDataOrder = requirement;
pFill->node.requireDataOrder = requirement;
return TSDB_CODE_SUCCESS;
}
static int32_t adjustSortDataRequirement(SSortLogicNode* pSort, EDataOrderLevel requirement) {
return TSDB_CODE_SUCCESS;
}
static int32_t adjustPartitionDataRequirement(SPartitionLogicNode* pPart, EDataOrderLevel requirement) {
if (DATA_ORDER_LEVEL_GLOBAL == requirement) {
return TSDB_CODE_PLAN_INTERNAL_ERROR;
}
pPart->node.resultDataOrder = requirement;
pPart->node.requireDataOrder = requirement;
return TSDB_CODE_SUCCESS;
}
static int32_t adjustIndefRowsDataRequirement(SIndefRowsFuncLogicNode* pIndef, EDataOrderLevel requirement) {
if (requirement <= pIndef->node.resultDataOrder) {
return TSDB_CODE_SUCCESS;
}
pIndef->node.resultDataOrder = requirement;
pIndef->node.requireDataOrder = requirement;
return TSDB_CODE_SUCCESS;
}
static int32_t adjustInterpDataRequirement(SInterpFuncLogicNode* pInterp, EDataOrderLevel requirement) {
if (requirement <= pInterp->node.requireDataOrder) {
return TSDB_CODE_SUCCESS;
}
pInterp->node.resultDataOrder = requirement;
pInterp->node.requireDataOrder = requirement;
return TSDB_CODE_SUCCESS;
}
int32_t adjustLogicNodeDataRequirement(SLogicNode* pNode, EDataOrderLevel requirement) {
int32_t code = TSDB_CODE_SUCCESS;
switch (nodeType(pNode)) {
case QUERY_NODE_LOGIC_PLAN_SCAN:
code = adjustScanDataRequirement((SScanLogicNode*)pNode, requirement);
break;
case QUERY_NODE_LOGIC_PLAN_JOIN:
code = adjustJoinDataRequirement((SJoinLogicNode*)pNode, requirement);
break;
case QUERY_NODE_LOGIC_PLAN_AGG:
code = adjustAggDataRequirement((SAggLogicNode*)pNode, requirement);
break;
case QUERY_NODE_LOGIC_PLAN_PROJECT:
code = adjustProjectDataRequirement((SProjectLogicNode*)pNode, requirement);
break;
case QUERY_NODE_LOGIC_PLAN_VNODE_MODIFY:
case QUERY_NODE_LOGIC_PLAN_EXCHANGE:
case QUERY_NODE_LOGIC_PLAN_MERGE:
break;
case QUERY_NODE_LOGIC_PLAN_WINDOW:
code = adjustWindowDataRequirement((SWindowLogicNode*)pNode, requirement);
break;
case QUERY_NODE_LOGIC_PLAN_FILL:
code = adjustFillDataRequirement((SFillLogicNode*)pNode, requirement);
break;
case QUERY_NODE_LOGIC_PLAN_SORT:
code = adjustSortDataRequirement((SSortLogicNode*)pNode, requirement);
break;
case QUERY_NODE_LOGIC_PLAN_PARTITION:
code = adjustPartitionDataRequirement((SPartitionLogicNode*)pNode, requirement);
break;
case QUERY_NODE_LOGIC_PLAN_INDEF_ROWS_FUNC:
code = adjustIndefRowsDataRequirement((SIndefRowsFuncLogicNode*)pNode, requirement);
break;
case QUERY_NODE_LOGIC_PLAN_INTERP_FUNC:
code = adjustInterpDataRequirement((SInterpFuncLogicNode*)pNode, requirement);
break;
default:
break;
}
if (TSDB_CODE_SUCCESS == code) {
SNode* pChild = NULL;
FOREACH(pChild, pNode->pChildren) {
code = adjustLogicNodeDataRequirement((SLogicNode*)pChild, pNode->requireDataOrder);
if (TSDB_CODE_SUCCESS != code) {
break;
}
}
}
return code;
}

6
source/libs/planner/test/planIntervalTest.cpp
View File

@ -38,9 +38,15 @@ TEST_F(PlanIntervalTest, fill) {
run("SELECT COUNT(*) FROM t1 WHERE ts > TIMESTAMP '2022-04-01 00:00:00' and ts < TIMESTAMP '2022-04-30 23:59:59' "
"INTERVAL(10s) FILL(LINEAR)");
run("SELECT COUNT(*) FROM st1 WHERE ts > TIMESTAMP '2022-04-01 00:00:00' and ts < TIMESTAMP '2022-04-30 23:59:59' "
"INTERVAL(10s) FILL(LINEAR)");
run("SELECT COUNT(*), SUM(c1) FROM t1 "
"WHERE ts > TIMESTAMP '2022-04-01 00:00:00' and ts < TIMESTAMP '2022-04-30 23:59:59' "
"INTERVAL(10s) FILL(VALUE, 10, 20)");
run("SELECT COUNT(*) FROM st1 WHERE ts > TIMESTAMP '2022-04-01 00:00:00' and ts < TIMESTAMP '2022-04-30 23:59:59' "
"PARTITION BY TBNAME interval(10s) fill(prev)");
}
TEST_F(PlanIntervalTest, selectFunc) {

20
source/libs/planner/test/planSubqueryTest.cpp
View File

@ -48,10 +48,28 @@ TEST_F(PlanSubqeuryTest, doubleGroupBy) {
"WHERE a > 100 GROUP BY b");
}
TEST_F(PlanSubqeuryTest, withSetOperator) {
TEST_F(PlanSubqeuryTest, innerSetOperator) {
useDb("root", "test");
run("SELECT c1 FROM (SELECT c1 FROM t1 UNION ALL SELECT c1 FROM t1)");
run("SELECT c1 FROM (SELECT c1 FROM t1 UNION SELECT c1 FROM t1)");
}
TEST_F(PlanSubqeuryTest, innerFill) {
useDb("root", "test");
run("SELECT cnt FROM (SELECT _WSTART ts, COUNT(*) cnt FROM t1 "
"WHERE ts > '2022-04-01 00:00:00' and ts < '2022-04-30 23:59:59' INTERVAL(10s) FILL(LINEAR)) "
"WHERE ts > '2022-04-06 00:00:00'");
}
TEST_F(PlanSubqeuryTest, outerInterval) {
useDb("root", "test");
run("SELECT COUNT(*) FROM (SELECT * FROM st1) INTERVAL(5s)");
run("SELECT COUNT(*) + SUM(c1) FROM (SELECT * FROM st1) INTERVAL(5s)");
run("SELECT COUNT(*) FROM (SELECT ts, TOP(c1, 10) FROM st1s1) INTERVAL(5s)");
}

13
source/libs/qcom/src/queryUtil.c
View File

@ -148,11 +148,12 @@ void destroySendMsgInfo(SMsgSendInfo* pMsgBody) {
taosMemoryFreeClear(pMsgBody);
}
int32_t asyncSendMsgToServerExt(void* pTransporter, SEpSet* epSet, int64_t* pTransporterId, const SMsgSendInfo* pInfo,
int32_t asyncSendMsgToServerExt(void* pTransporter, SEpSet* epSet, int64_t* pTransporterId, SMsgSendInfo* pInfo,
bool persistHandle, void* rpcCtx) {
char* pMsg = rpcMallocCont(pInfo->msgInfo.len);
if (NULL == pMsg) {
qError("0x%" PRIx64 " msg:%s malloc failed", pInfo->requestId, TMSG_INFO(pInfo->msgType));
destroySendMsgInfo(pInfo);
terrno = TSDB_CODE_TSC_OUT_OF_MEMORY;
return terrno;
}
@ -167,13 +168,15 @@ int32_t asyncSendMsgToServerExt(void* pTransporter, SEpSet* epSet, int64_t* pTra
.info.persistHandle = persistHandle,
.code = 0
};
assert(pInfo->fp != NULL);
TRACE_SET_ROOTID(&rpcMsg.info.traceId, pInfo->requestId);
rpcSendRequestWithCtx(pTransporter, epSet, &rpcMsg, pTransporterId, rpcCtx);
return TSDB_CODE_SUCCESS;
int code = rpcSendRequestWithCtx(pTransporter, epSet, &rpcMsg, pTransporterId, rpcCtx);
if (code) {
destroySendMsgInfo(pInfo);
}
return code;
}
int32_t asyncSendMsgToServer(void* pTransporter, SEpSet* epSet, int64_t* pTransporterId, const SMsgSendInfo* pInfo) {
int32_t asyncSendMsgToServer(void* pTransporter, SEpSet* epSet, int64_t* pTransporterId, SMsgSendInfo* pInfo) {
return asyncSendMsgToServerExt(pTransporter, epSet, pTransporterId, pInfo, false, NULL);
}

20
source/libs/scheduler/src/schRemote.c
View File

@ -509,7 +509,7 @@ int32_t schGenerateCallBackInfo(SSchJob *pJob, SSchTask *pTask, void *msg, uint3
SMsgSendInfo *msgSendInfo = taosMemoryCalloc(1, sizeof(SMsgSendInfo));
if (NULL == msgSendInfo) {
SCH_TASK_ELOG("calloc %d failed", (int32_t)sizeof(SMsgSendInfo));
SCH_ERR_RET(TSDB_CODE_QRY_OUT_OF_MEMORY);
SCH_ERR_JRET(TSDB_CODE_QRY_OUT_OF_MEMORY);
}
msgSendInfo->paramFreeFp = taosMemoryFree;
@ -535,8 +535,12 @@ int32_t schGenerateCallBackInfo(SSchJob *pJob, SSchTask *pTask, void *msg, uint3
_return:
destroySendMsgInfo(msgSendInfo);
if (msgSendInfo) {
destroySendMsgInfo(msgSendInfo);
}
taosMemoryFree(msg);
SCH_RET(code);
}
@ -843,6 +847,7 @@ int32_t schAsyncSendMsg(SSchJob *pJob, SSchTask *pTask, SSchTrans *trans, SQuery
int64_t transporterId = 0;
code = asyncSendMsgToServerExt(trans->pTrans, epSet, &transporterId, pMsgSendInfo, persistHandle, ctx);
pMsgSendInfo = NULL;
if (code) {
SCH_ERR_JRET(code);
}
@ -919,7 +924,9 @@ int32_t schBuildAndSendHbMsg(SQueryNodeEpId *nodeEpId, SArray *taskAction) {
addr.epSet.numOfEps = 1;
memcpy(&addr.epSet.eps[0], &nodeEpId->ep, sizeof(nodeEpId->ep));
SCH_ERR_JRET(schAsyncSendMsg(NULL, NULL, &trans, &addr, msgType, msg, msgSize, true, &rpcCtx));
code = schAsyncSendMsg(NULL, NULL, &trans, &addr, msgType, msg, msgSize, true, &rpcCtx);
msg = NULL;
SCH_ERR_JRET(code);
return TSDB_CODE_SUCCESS;
@ -1087,9 +1094,10 @@ int32_t schBuildAndSendMsg(SSchJob *pJob, SSchTask *pTask, SQueryNodeAddr *addr,
}
SSchTrans trans = {.pTrans = pJob->conn.pTrans, .pHandle = SCH_GET_TASK_HANDLE(pTask)};
SCH_ERR_JRET(
schAsyncSendMsg(pJob, pTask, &trans, addr, msgType, msg, msgSize, persistHandle, (rpcCtx.args ? &rpcCtx : NULL)));
schAsyncSendMsg(pJob, pTask, &trans, addr, msgType, msg, msgSize, persistHandle, (rpcCtx.args ? &rpcCtx : NULL));
msg = NULL;
SCH_ERR_JRET(code);
if (msgType == TDMT_SCH_QUERY || msgType == TDMT_SCH_MERGE_QUERY) {
SCH_ERR_RET(schAppendTaskExecNode(pJob, pTask, addr, pTask->execId));
}

16
source/libs/scheduler/src/schTask.c
View File

@ -102,14 +102,14 @@ int32_t schRecordTaskSucceedNode(SSchJob *pJob, SSchTask *pTask) {
}
int32_t schAppendTaskExecNode(SSchJob *pJob, SSchTask *pTask, SQueryNodeAddr *addr, int32_t execId) {
SSchNodeInfo nodeInfo = {.addr = *addr, .handle = NULL};
SSchNodeInfo nodeInfo = {.addr = *addr, .handle = SCH_GET_TASK_HANDLE(pTask)};
if (taosHashPut(pTask->execNodes, &execId, sizeof(execId), &nodeInfo, sizeof(nodeInfo))) {
SCH_TASK_ELOG("taosHashPut nodeInfo to execNodes failed, errno:%d", errno);
SCH_ERR_RET(TSDB_CODE_QRY_OUT_OF_MEMORY);
}
SCH_TASK_DLOG("task execNode added, execId:%d", execId);
SCH_TASK_DLOG("task execNode added, execId:%d, handle:%p", execId, nodeInfo.handle);
return TSDB_CODE_SUCCESS;
}
@ -752,12 +752,18 @@ void schDropTaskOnExecNode(SSchJob *pJob, SSchTask *pTask) {
return;
}
int32_t i = 0;
SSchNodeInfo *nodeInfo = taosHashIterate(pTask->execNodes, NULL);
while (nodeInfo) {
SCH_SET_TASK_HANDLE(pTask, nodeInfo->handle);
schBuildAndSendMsg(pJob, pTask, &nodeInfo->addr, TDMT_SCH_DROP_TASK);
if (nodeInfo->handle) {
SCH_SET_TASK_HANDLE(pTask, nodeInfo->handle);
schBuildAndSendMsg(pJob, pTask, &nodeInfo->addr, TDMT_SCH_DROP_TASK);
SCH_TASK_DLOG("start to drop task's %dth execNode", i);
} else {
SCH_TASK_DLOG("no need to drop task %dth execNode", i);
}
++i;
nodeInfo = taosHashIterate(pTask->execNodes, nodeInfo);
}

2
source/libs/stream/src/streamData.c
View File

@ -34,6 +34,7 @@ int32_t streamDispatchReqToData(const SStreamDispatchReq* pReq, SStreamDataBlock
// TODO: refactor
pDataBlock->info.window.skey = be64toh(pRetrieve->skey);
pDataBlock->info.window.ekey = be64toh(pRetrieve->ekey);
pDataBlock->info.version = be64toh(pRetrieve->version);
pDataBlock->info.type = pRetrieve->streamBlockType;
pDataBlock->info.childId = pReq->upstreamChildId;
@ -54,6 +55,7 @@ int32_t streamRetrieveReqToData(const SStreamRetrieveReq* pReq, SStreamDataBlock
// TODO: refactor
pDataBlock->info.window.skey = be64toh(pRetrieve->skey);
pDataBlock->info.window.ekey = be64toh(pRetrieve->ekey);
pDataBlock->info.version = be64toh(pRetrieve->version);
pDataBlock->info.type = pRetrieve->streamBlockType;

Some files were not shown because too many files have changed in this diff Show More