转自:https://severalnines.com/blog/creating-new-modules-using-postgresql-create-extension

Extensibility is one of the most powerful feature in PostgreSQL. You can add new functionality for a particular use case by using contrib module and install it using CREATE EXTENSION.

In this section, we are going to learn how to create a simple contrib module and how to use its functionality in PostgreSQL.

Extension Files

To be able to run the CREATE EXTENSION command in your database, your extension must need at least two files:

  1. Control file
    The file format must be extension_name.control, which tells the basics about extension to PostgreSQL, and must be placed in the installation’s SHAREDIR/extension directory.
  2. SQL script file
    The file in the format extension--version.sql contains the functions that you would like to add.

The file format of the control file in the extension is same as postgresql.conf file, namely a list of parameter_name = value assignments, one per line.

Example

Please check the below complete example of an SQL-only extension, create Oracle compatible NVL function in PostgreSQL. There are many cases but here we can consider only one case for example.

The SQL script file nvlfunc--1.0.sql looks like this...

Nvlfunc--1.0.sql file:

1
2
3
4
5
6
7
--complain if script is sourced in psql, rather than via CREATE EXTENSION
\echo Use "CREATE EXTENSION nvlfunc" to load this file. \quit
 
CREATE OR REPLACE FUNCTION public.NVL(SMALLINT,SMALLINT)
RETURNS SMALLINT AS $$
SELECT COALESCE($1,$2);
$$ LANGUAGE SQL IMMUTABLE;

The control file nvlfunc looks like this...

Nvlfunc.conntrol file:

1
2
3
4
5
# nvlfunc extension
comment = 'Oracle compatible nvl function'
default_version = '1.0'
module_pathname = '$libdir/nvlfunc'
relocatable = false

While you hardly need a makefile to install these files into the correct directory, you could use a Makefile containing this:

Makefile:

1
2
3
4
5
6
EXTENSION = nvlfunc
DATA = nvlfunc--1.0.sql
 
PG_CONFIG = pg_config
PGXS := $(shell $(PG_CONFIG) --pgxs)
include $(PGXS)

If you have implemented the function using ‘C’ language then you need to add the file in the makefile.

Installation

The command make install will install the control file and sql script file into the correct directory as reported by pg_config.

Once the files are installed, use the CREATE EXTENSION command to load the objects into any particular database in PostgreSQL.

Please check the following steps to install the nvlfunc extension and you can also add this file in your extension directory:

INSTALL.nvlfunc file:

1
2
3
4
5
6
7
8
9
10
11
12
This module is a PostgreSQL extension which provides the Oracle compatible nvl function feature.
Use below command in source directory of nvlfunc to install the module.
make install
Then use the below command to create extension nvlfunc in database.
CREATE EXTENSION nvlfunc;
Use the below command to remove the nvlfunc extension from database.
DROP EXTENSION nvlfunc;
Use below command in source directory of nvlfunc to uninstall the module.
make uninstall
Note:
This extension module requires PostgreSQL 9.1 or later because CREATE EXTENSION
feature is available in PostgreSQL 9.1 and later version.

PostgreSQL extensions must be installed in your database before you can use their functionality. To install a particular extension, run the CREATE EXTENSION command from psql to load the packaged objects into the database.

Testing

Once you create the extension, it is recommended to create a test case for that extension so that when you are installing this module in a different PostgreSQL version, you can then check whether all the test cases are working as expected or not.

This is an optional step but you can create it. The test cases look like this:

sql/nvlfunc.sql file:

1
SELECT NVL(NULL::SMALLINT, 11::SMALLINT);

Then you can add expected output in another file. The expected output file looks like this:

expected/nvlfunc.out file:

1
2
3
4
5
SELECT NVL(NULL::SMALLINT, 11::SMALLINT);
 nvl
-----
  11
(1 row)

This is just one example, that’s why only one test case is added. But when you are creating a new extension, you can add more test cases.

Nvlfunc extension directory structure:

1
2
3
4
# mkdir nvlfunc
# cd nvlfunc
# ls
Makefile         nvlfunc.control    nvlfunc--1.0.sql    sql     expected    INSTALL.nvlfunc    README
Related resources

Pros

  1. Easy to extend the PostgreSQL functionality
  2. Very easy to create and install the extension
  3. Easy to test for regressions on different versions of PostgreSQL

Cons

  1. There is no special cons but test the feature you are adding in the extension before you use it.
 
 
 
 

How to Create an PostgreSQL Extension的更多相关文章

  1. use Swig to create C/C++ extension for Python

    SWIG is a software development tool that simplifies the task of interfacing different languages to C ...

  2. 开发一个简单的postgresql extension

      主要是学习如何编写一个简单的pg extension,参考https://severalnines.com/blog/creating-new-modules-using-postgresql-c ...

  3. 使用fpm 方便快速生成postgresql extension分发包

    fpm 是一个不错,而且强大的rpm.deb,系统启动服务工具包,我们可以用来快速的生成专业的软件分发包 演示一个pg 扩展包分发包的生成(rpm 以及deb) 安装fpm sudo gem inst ...

  4. [Tools] Create a Chrome Extension

    Creating a Chrome extension requires a manifest.json file which defines how your extension will beha ...

  5. [官网]CREATE EXTENSION PostGreSQL 创建函数的方法

    CREATE EXTENSION https://www.postgresql.org/docs/current/sql-createextension.html CREATE EXTENSION — ...

  6. How to debug PostgreSQL function with pgAdminIII

    How to debug plpgsql with pgAdminIII [root@localhost soft_bak]# git clone git://git.postgresql.org/g ...

  7. ubuntu安装postgresql与postgis

    版本信息 ubuntu    14.04.1LTS postgresql   9.3.5 postgis       2.1.2 今天尝试着安装了postgis 2.1.2,(较简便的包安装,不是源码 ...

  8. 使用rpm 打包开发的postgres extension

      环境准备 安装依赖包 rpmdevtools rpm-build yum install -y rpm-build rpmdevtools 初始化rpm pacakge 项目 主要是rpm 打包的 ...

  9. PostgreSQL通过mysql_fdw访问MySQL数据库

    Mysql与PostgreSQL的安装过程省略. 为简便起见,把MySQL和PostgreSQL都安装在一个机器上,然后在此机器上(准确地说是在PostgreSQL运行的机器上)安装mysql_fdw ...

随机推荐

  1. 安装完MySQL数据库设置密码

    对于windows平台来说安装完MySQL数据库后,系统就已经默认生成了许可表和账户,你不需要像在Unix平台上那样执行 mysql_install_db脚本来生成帐户和相应权限许可表.但是如果不是用 ...

  2. HADOOP HA 踩坑 - 所有 namenode 都是standby

    报错: 无明显报错 状况: 所有namenode都是standby,即ZK服务未生效 尝试一:手动强制转化某个namenode为active 操作:在某台namenode上,执行 hdfs haadm ...

  3. 《深入理解Nginx:模块开发与架构解析》读书笔记

    1.nginx的特点:快.扩展性强.可靠性强.内存低消耗.支持高并发.热部署.开源免费 2.nginx由master进程来管理多个(CPU数)worker进程 3.配置按功能分,有4类: 1)用于调试 ...

  4. Spark面试相关

    Spark Core面试篇01 随着Spark技术在企业中应用越来越广泛,Spark成为大数据开发必须掌握的技能.前期分享了很多关于Spark的学习视频和文章,为了进一步巩固和掌握Spark,在原有s ...

  5. (17)线程队列---queue LifoQueue PriorityQueue

    线程常用队列有: queue LifoQueue PriorityQueue 语法: 大致和进程队列语法一致 put 往队列当中放值,超过队列长度,直接加阻塞 get 如果获取不到加阻塞 put_no ...

  6. RabbitMQ(1) - win+rabbitMQ

    rabbitMQ是一个在AMQP协议标准基础上完整的,可服用的企业消息系统.它遵循Mozilla Public License开源协议,采用 Erlang 实现的工业级的消息队列(MQ)服务器,Rab ...

  7. js的event事件对象汇总

    JavaScript事件对象是浏览器默认传入的,但是对于浏览器的兼容问题,我们需要对事件对象进行兼容.但是jQuery已经帮我们解决了所有兼容性的问题,并且给我们添加了很多有用的方法.已经是比较历史的 ...

  8. mysql(5.5)安装后忘记密码

    查看mysql安装的路径

  9. java 常见语法,但是发现switch等基础,常见面试套路不会了,待补充

    1,面向对象,万物皆对象,面向对象的三个基本特性:封 装,继承,多态. 面向对象的程序都是由类组成,类封装了对象的属性和行为. 封装:隐藏对象的属性和实现细节,仅对外提供接口. 继承:实现代码复用,将 ...

  10. <转载> 从算法上解读自动驾驶是如何实现的?

     科技新闻小鹏汽车2016-03-28 10:42 [摘要]车辆路径规划问题中路网模型.路径规划算法和交通信息的智能预测为关键点. 由于驾驶员的驾驶工作繁重,同时随着汽车拥有量的增加,非职业驾驶员的数 ...