postgres

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15

podman run --name postgres \
-p 5432:5432 \
-e POSTGRES_PASSWORD=password0 \
-e PGDATA=/var/lib/postgresql/data/pgdata \
-v postgres-data:/var/lib/postgresql/data \
-d postgres:14.5

psql -h 127.0.0.1 -p 5432 -U postgres
# password: password0

# 重启服务
su -l postgres -c '/opt/pg9.6/bin/pg_ctl -D /mnt/pgdata start'

# 重载配置
su -l postgres -c '/opt/pg9.6/bin/pg_ctl -D /mnt/pgdata reload'

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36

# 查看配置文件路径, 切换到 postgres 用户执行
psql -c "show config_file"

# 查看版本
select version();

pacman -S postgresql
psql -h 127.0.0.1 -p 5432 -d database0 -U user0

# create database
create database database0;

# create table
create table test(id int, c1 int);
create table table0(field0 json);

# delete table
DROP TABLE table0;
# 查看字段类型
select column_name, data_type from information_schema.columns where table_name='table0';

\l 或 \list meta-command 列出所有数据库
sudo -u postgres psql -c "\l"
用 \c + 数据库名 来进入数据库：
\dt 列出所有数据库表：

# 查看表结构
\d table0

# Turn off printing of column names and result row count footers, etc. This is equivalent to \t or \pset tuples_only.
\t tuples only on/off, tuples only on 的时候 select 语句的输出不带 header

\h
\?
select * length( "abc"::TEXT)
insert into test select generate_series(1,10000), random()*10;

1
2
3
4
5

alter table foo
    rename to bar;

alter table foo
    add time0 timestamptz;

1
2
3
4

create table table0(
  field0 json,
  create_time             timestamp with time zone default now(),
  );

1

PGPASSWORD=postgres psql -h 127.0.0.1 -p 5432 -d database0 -U user0  --command 'select version();'

1

PGPASSWORD=postgres psql -h 127.0.0.1 -p 5432 -d database0 -U user0 -f /path/to/foo.sql

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

--查询主键名称
SELECT
    pg_constraint.conname AS pk_name
FROM
    pg_constraint
INNER JOIN pg_class ON pg_constraint.conrelid = pg_class.oid
WHERE
    pg_class.relname = 'table_name_0'
AND pg_constraint.contype = 'p';
--查询主键的详细信息
SELECT
    pg_constraint.conname AS pk_name,
    pg_attribute.attname AS colname,
    pg_type.typname AS typename
FROM
    pg_constraint
INNER JOIN pg_class ON pg_constraint.conrelid = pg_class.oid
INNER JOIN pg_attribute ON pg_attribute.attrelid = pg_class.oid
AND pg_attribute.attnum = pg_constraint.conkey [ 1 ]
INNER JOIN pg_type ON pg_type.oid = pg_attribute.atttypid
WHERE
    pg_class.relname = 'table_name_0'
AND pg_constraint.contype = 'p';

1
2
3
4

SELECT to_char(NOW(), 'YYYY-MM-DD HH24:MI:SS')
-- 时区
to_char(NOW() AT TIME ZONE 'Asia/Dubai','YYYY-MM-DD HH24:MI:SS')

1
2
3
4

select now() + interval '1 days'; 
select now() + interval '1 month'; 
select now() + interval '1 years'; 
select NOW(), NOW() - interval '1 hours 4 minutes';

1

string||string

1
2

-- 单引号转义 ''
UPDATE user SET username = 'Peter''s Name' WHERE id = 1;

1
2
3
4
5
6

SELECT to_number('12345', '9999999999999999999')//12345
SELECT to_number('12345', '99999')//12345
SELECT to_number(''||12345, '9999')//1234，由于模式是4位，结果忽略最后一位；
SELECT to_number('    12345', '9999999999999999999')//12345
SELECT to_number('  ab  ,1,2a3,4b5', '9999999999999999999')//12345，会忽略所有字符串中非数字字符

 1
 2
 3
 4
 5
 6
 7
 8
 9
10

# -h, host 127.0.0.1
# -p, port 5432
# -t, table: table0, 不加 -t 参数时会导出所有表结构
# -s, 不导出数据
# database: database0
pg_dump -h 127.0.0.1 -p 5432 -t table0 -U postgres database0 > foo.sql
pg_dump -h 127.0.0.1 -p 5432 -s -t table0 -U postgres database0 > foo.sql

# 导出并压缩
pg_dump -d db_name | gzip > db.gz

1

psql -h 127.0.0.1 -p 5432 -t table0 -U postgres -d database0 -f foo.sql

1
2

CREATE SEQUENCE seq_0 START 1;
CREATE SEQUENCE seq_0 INCREMENT 1 MINVALUE 1 START 1 CACHE 1;  

1
2
3
4
5

\d [tablename]
"table_name_id_fkey" FOREIGN KEY (id) REFERENCES other_table(id) ....
-- 然后使用下面的命令删除外键：

ALTER TABLE [tablename] DROP CONSTRAINT table_name_id_fkey;

 1
 2
 3
 4
 5
 6
 7
 8
 9
10

SELECT
     tc.constraint_name, tc.table_name, kcu.column_name, 
     ccu.table_name AS foreign_table_name,
     ccu.column_name AS foreign_column_name,
     tc.is_deferrable,tc.initially_deferred
FROM
     information_schema.table_constraints AS tc 
     JOIN information_schema.key_column_usage AS kcu ON tc.constraint_name = kcu.constraint_name
     JOIN information_schema.constraint_column_usage AS ccu ON ccu.constraint_name = tc.constraint_name
WHERE constraint_type = 'FOREIGN KEY' AND tc.table_name = 'table0';

1
2
3
4
5

名字                        别名             描述
character varying [(n)]  varchar [ (n) ]  可变长字符串
character [(n)]          char [ (n) ]     定长字符串
timestamp                                   SQL标准要求仅仅将timestamp类型等于timestamp without time zone 类型
timestamp with time zone   TIMESTAMPTZ       PostgreSQL遵守这个行为。timestamptz 作为 timestamp with time zone 的缩写被接受；这是PostgreSQL 的一个扩展。

1
2
3
4
5

show timezone;
select * from pg_timezone_names where abbrev='+04';
set time zone "Asia/Dubai";

select now();

1
2

substring(string [from <str_pos>] [for <ext_char>])
-- str_pos 

1

select * from pg_tables where tablename = 'my_tbl';

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24

-- 查看锁
SELECT locker.pid,
       pc.relname,
       locker.mode,
       locker_act.application_name,
       least(query_start, xact_start)                        start_time,
       locker_act.state,
       CASE
           WHEN granted = 'f' THEN
               'wait_lock'
           WHEN granted = 't' THEN
               'get_lock'
           END                                               lock_satus,
       current_timestamp - least(query_start, xact_start) AS runtime,
       locker_act.query
FROM pg_locks locker,
     pg_stat_activity locker_act,
     pg_class pc
WHERE locker.pid = locker_act.pid
  AND NOT locker.pid = pg_backend_pid()
  AND application_name <> 'pg_statsinfod'
  AND locker.relation = pc.oid
  AND pc.reltype <> 0 --and pc.relname='t'
ORDER BY runtime desc;

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16

--查看 PostgreSQL 正在执行的SQL  
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,
     pg_stat_activity pa
WHERE current_query <> '<IDLE>'
  and procpid <> pg_backend_pid()
  and pa.pid = s.procpid
  and pa.state <> 'idle'
ORDER BY lap DESC;

1
2

# 查找是否有 waiting  
ps -ef|grep postgres | grep wait  

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14

--查看当前库表和索引的的大小并排序显示前20条  
SELECT nspname,
       relname,
       relkind                                       as "type",
       pg_size_pretty(pg_table_size(C.oid))          AS size,
       pg_size_pretty(pg_indexes_size(C.oid))        AS idxsize,
       pg_size_pretty(pg_total_relation_size(C.oid)) as "total"
FROM pg_class C
         LEFT JOIN pg_namespace N ON (N.oid = C.relnamespace)
WHERE nspname NOT IN ('pg_catalog', 'information_schema')
  AND nspname !~ '^pg_toast'
  AND relkind IN ('r', 'i')
ORDER BY pg_total_relation_size(C.oid) DESC
LIMIT 20;

 1
 2
 3
 4
 5
 6
 7
 8
 9
10

-- 最大连接数
show max_connections;
-- 当前连接
select * from pg_stat_activity;
select datname from pg_stat_activity group by datname;
select state from pg_stat_activity group by state;
select * from pg_stat_activity where datname='database0' limit 1;
select pid, query_start,query from pg_stat_activity where datname='database0' and state='idle';
-- 释放空闲连接
SELECT pg_terminate_backend(pid) FROM pg_stat_activity WHERE state='idle';

1
2

copy (select now()) to '/tmp/foo.csv' with csv header;
psql --dbname=database0 --host=127.0.0.1 --username=user0 -c "COPY (select now();) TO STDOUT with csv header" > /tmp/foo.csv

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26

# 默认值: off, 改成 on 之后重启 postgresql 就开始写日志了, 日志目录会是 /var/lib/postgresql/data/pgdata/log/
logging_collector = on

# none, ddl, mod, all ---- 控制记录哪些 SQL 语句。none 不记录，ddl 记录所有数据定义命令，比如 CREATE, ALTER, 和 DROP 语句。mod 记录所有 ddl 语句, 加上数据修改语句 INSERT,UPDATE等,all记录所有执行的语句，将此配置设置为all可跟踪整个数据库执行的SQL语句。
log_statement = 'all'

# 日志目录名, 默认值: log
# log_directory = 'log'

# 默认文件名
log_filename = 'postgresql-%Y-%m-%d_%H%M%S.log'

log_rotation_age = 1d ----  单个日志文件的生存期，默认 1 天，在日志文件大小没有达到 log_rotation_size 时，一天只生成一个日志文件
log_rotation_size = 10MB  ---- 单个日志文件的大小，如果时间没有超过 log_rotation_age，一个日志文件最大只能到 10M，否则将新生成一个日志文件。
log_truncate_on_rotation = off ---- 当日志文件已存在时，该配置如果为off，新生成的日志将在文件尾部追加，如果为on，则会覆盖原来的日志。
log_duration = off ---- 记录每条SQL语句执行完成消耗的时间，将此配置设置为 on, 用于统计哪些 SQL 语句耗时较长。

10.log_min_duration_statement = -1 # -1 is disabled, 0 logs all statements and their durations, > 0 logs only statements running at least this number of milliseconds

-1表示不可用，0将记录所有SQL语句和它们的耗时，>0只记录那些耗时超过（或等于）这个值（ms）的SQL语句。个人更喜欢使用该配置来跟踪那些耗时较长，可能存在性能问题的SQL语句。虽然使用log_statement和log_duration也能够统计SQL语句及耗时，但是SQL语句和耗时统计结果可能相差很多行，或在不同的文件中，但是log_min_duration_statement会将SQL语句和耗时在同一行记录，更方便阅读。

11.log_connections = off ----是否记录连接日志
12.log_disconnections = off ---- 是否记录连接断开日志

13.log_line_prefix = '%m %p %u %d %r ' ---- 日志输出格式（%m,%p实际意义配置文件中有解释）,可根据自己需要设置（能够记录时间，用户名称，数据库名称，客户端IP和端口，方便定位问题）
14.log_timezone = 'Asia/Shanghai' ---- 日志时区，最好和服务器设置同一个时区，方便问题定位

1

select array_length(regexp_split_to_array(config,'"id":'),1)-1 from xxx;

1
2
3
4
5
6
7

### 操作系统上查看 WAL 发送进程或 WAL 接收进程
ps -ef | grep "wal" | grep -v "grep"
# 主库会有 postgres: walwriter, postgres: walsender 进程
# 从库只有 postgres: walreceiver 进程

# 通过 pg_controldata 命令查看数据库控制信息，内容包含 WAL 日志信息、checkpoint、数据块等信息，通过 Databasecluster state 信息可判断是主库还是备库
pg_controldata | grep cluster

1
2
3
4

-- psql 连接到 DB 之后, 可以这样判断
SELECT pg_is_in_recovery();
-- You can use pg_is_in_recovery() which returns True if recovery is still in progress(so the server is running in standby mode). Check the System Administration Functions for further informations.
-- 如果返回 t 说明是备库，返回 f 是主库

1
2
3
4
5

create table test(id int, c1 int);  
insert into test select generate_series(1,10000), random()*10;
select * from test limit 10;
select c1,count(*)  from test group by c1;
select c1,count(*), count(*) filter (where id<1000) from test group by c1;