postgresql基本语法

启动和停止数据库：

postgres -D /export/pgsql/data/ > /export/pgsql/log/pg_server.log 2>&1 &

pg_ctl stop -D /export/pgsql/data/

1.登录

sudo su - postgres #需要使用postgres用户

psql -h localhost -p 5432 #-h指定主机名，-p指定监听端口号

\c testdb testch

show search_path; --查看当前的查找路径，schema

set search_path= "$user", public,wala;

\d --查看wala schema下的所有表、视图、序列等数据库对象

\d c_comment --describe database objects' DEFINITION

-- 显示当前默认schema

show search_path;

-- 设置默认schema,退出后再登录，就失效了

set search_path='scrm';

2.查看当前正在执行的语句

\d pg_stat_activity --\d=desc tablename

select datname,usename,client_hostname,client_port,application_name,query

from pg_stat_activity

where query!='<IDLE>';

select pid,usename,client_hostname,state,query from pg_stat_activity;

1.查看数据库软件版本

select version();

2.查看服务器运行时间

select date_trunc('second',current_timestamp - pg_postmaster_start_time()) as uptime;

3.查看数据库大小

select datname, pg_database_size(datname)/1024/1024 "dbsize M" from pg_database;

查看表的大小

select pg_size_pretty(pg_relation_size('test01'));

查看表的索引的大小

select pg_size_pretty(pg_indexes_size('test01'));

查看表空间大小

select pg_size_pretty(pg_tablespace_size('pg_global'));

select pg_size_pretty(pg_tablespace_size('pg_default'));

4.切换日志

设置日志启动方式：

log_truncate_on_rotation = on(postgres.conf)

select pg_rotate_logfile();

切换WAL日志

select pg_switch_xlog();

手动checkpoint

checkpoint;

4.根据后台进程号，查看当前执行的sql语句：

select *

from (

select pg_stat_get_backend_pid(s.backendid) as procid,

pg_stat_get_backend_activity(s.backendid) as current_query

from (

select pg_stat_get_backend_idset() as backendid

)

as s

) as querystring

where procid=xxxx;

找出长时间运行的SQL

select pid,usename,query_start,query from pg_stat_activity;

SELECT procpid, start, now() - start AS lap,

current_query

FROM

(SELECT

backendid,

pg_stat_get_backend_pid(S.backendid) AS procpid,

pg_stat_get_backend_activity_start(S.backendid) AS start,

pg_stat_get_backend_activity(S.backendid) AS current_query

FROM

(SELECT pg_stat_get_backend_idset() AS backendid) AS S

) AS S

WHERE

current_query <> '<IDLE>'

ORDER BY

lap DESC;

5.杀死某个session,杀死长时间运行的SQL

select pg_cancel_backend($SID);

select pg_terminate_backend($SID);

mkdir -p /u01/app/postgres/data/tbs_gpsp on dbserver3,dbserver4

---/opt/postgres/data/pg_tbsdir/tbs_gpsp

CREATE ROLE gpsp LOGIN PASSWORD 'PASSWORD' NOSUPERUSER INHERIT NOCREATEDB NOCREATEROLE NOREPLICATION;

CREATE TABLESPACE tbs_gpsp OWNER gpsp LOCATION '/u01/app/postgres/data/tbs_gpsp';

--CREATE TABLESPACE tbs_gpsp OWNER gpsp LOCATION '/opt/postgres/data/pg_tbsdir/tbs_gpsp';

CREATE DATABASE gpsp WITH OWNER = gpsp ENCODING = 'UTF8' TABLESPACE = tbs_gpsp LC_COLLATE = 'en_US.UTF8' LC_CTYPE = 'en_US.UTF8' CONNECTION LIMIT = -1;

CREATE SCHEMA gpsp AUTHORIZATION gpsp;

--GRANT ALL ON SCHEMA gpsp TO gpsp;

CREATE TABLE gpsp.joblock(

jobname character varying(80) NOT NULL,

firetime character varying(20),

ip character varying(15),

status character varying(10),

nextfire character varying(20) NOT NULL,

CONSTRAINT pk_joblock PRIMARY KEY (jobname , nextfire)

);

ALTER TABLE gpsp.joblock OWNER TO gpsp;

序列的采用：

nextval('gsp.seq_check_bill')

1.PG序列的创建

方法a：直接在表中指定字段类型是 SERIAL 类型就一了

CREATE TABLE person (id SERIAL,name TEXT);

方法b：先创建序列名称，然后在新建的表中列属性指定序列就可以了，该列需INT类型列

CREATE SEQUENCE person_id_seq;

CREATE TABLE person (id INT4 NOT NULL DEFAULT nextval('person_id_seq'),name TEXT );

2.PG序列的应用

查看当前会话的序列值

SELECT currval('person_id_seq');

查看下一个序列值

SELECT nextval('person_id_seq') ;

查看全局的最后一个序列值

select last_value from person_id_seq;

重置序列值

select setval('person_id_seq',8);

3.序列删除当有表字段应用到PG序列时不能直接删除

drop table table_name;

drop sequence seq_name;

索引重建：

create index CONCURRENTLY idx_id on table (id);

--重建

遇到上述失效的索引重建时两个办法，一个是drop index index_name，然后再执行create index concurrently。还有一个是执行reindex index_name命令，但是后者不支持concurrent参数。

表导出导入

导出表 employee 到默认输出 STDOUT

COPY employee TO STDOUT;

导出表 employee 到 sql 文件：

COPY employee TO '/home/smallfish/employee.sql';

从文件导入数据：

COPY employeenew FROM '/home/smallfish/employee.sql';

导入文件数据，指定间隔符为 | ：

COPY employeenew FROM '/home/smallfish/employee.sql' DELIMITER |;

导出为CVS格式：

COPY employee (name, email) TO '/home/smallfish/employee.csv' CSV HEADER;

create table t (id int,name varchar(20));

insert into t values (1,'chen');

insert into t values (2,'chen123');

Postgresql数据的导入和导出，以及copy命令介绍

如何导出PostgreSQL数据库中的数据：

导出scrm库：

pg_dump -U postgres -f dump.sql scrm;

具体person表：

pg_dump -U postgres -t person -f scrm_person_dump.sql scrm

导入数据时首先创建数据库再用psql导入：

drop database gpsp;

create database gpsp;

psql -d scrm -U postgres -f dump.sql

psql -d scrm -U postgres -f scrm_person_dump.sql

把数据按照自己所想的方式导出，强大的copy命令：

echo "copy students to? stdout DELIMITER '|'"|psql school|head

(students为表名，school为库名，各个字段以|分隔)

echo 'copy (select * from students order by age limit 10) to stdout;' | psql school

导出为CVS格式：

COPY employee (name, email) TO '/home/smallfish/employee.csv' CSV HEADER;

导出csv文件

COPY gpsp.t2 (id) TO '/home/postgres/android2.csv' CSV HEADER;

导入csv文件:要以管理员身份导入

create table t2 (id);

psql

\c scrm

COPY scrm.person (id) from '/home/postgres/person.csv' with CSV HEADER;

备注: 当需要导出表数据到文件时， copy 命令需要以超级用户执行。而 \copy 元子命令不需要。

由于 copy 命令始终是到数据库服务端找文件，当以文件形式导入导出数据时需以

超级用户执行，权限要求很高，适合数据库管理员操作;而 \copy 命令可在客户端执行导

入客户端的数据文件，权限要求没那么高，适合开发人员，测试人员使用，因为生产库

的权限掌握在 DBA 手中。

备注：由于 copy 命令会到服务端找文件，所以找不到。

set search_path=scrm;

\d

导入csv:

psql -h 103.20.250.36 -p 5432 -U infocenter -d report

create table tempa (id bigint);

alter table infocenter.tempa owner to infocenter;

\copy infocenter.tempa (id) from '/home/postgres/apple.csv' with CSV HEADER;