当前位置 博文首页 > 利用python为PostgreSQL的表自动添加分区
PostgreSQL引进“分区”表特性，解放了之前采用“表继承”+“触发器”来实现分区表的繁琐、低效。而添加分区，都是手动执行SQL。
演示目的：利用python来为PostgreSQL的表自动添加分区。
python版本：python3+
pip3 install psycopg2
database.ini 文件：记录数据库连接参数
[adsas] host=192.168.1.201 database=adsas user=adsas password=adsas123 port=5432 [test] host=192.168.1.202 database=adsas user=adsas password=adsas123 port=5432
config.py 文件：下面的 config() 函数读取 database.ini 文件并返回连接参数。config() 函数位于config.py文件中
#!/usr/bin/python3
from configparser import ConfigParser
def config(section ,filename='database.ini'):
# create a parser
parser = ConfigParser()
# read config file
parser.read(filename)
# get section, default to postgresql
db = {}
if parser.has_section(section):
params = parser.items(section)
for param in params:
db[param[0]] = param[1]
else:
raise Exception('Section {0} not found in the {1} file'.format(section, filename))
return db
pg_add_partition_table.py 文件：其中 create_table函数是创建子表SQL。其中参数
| 参数名 | 含义 |
|---|---|
| db | 指向数据库 |
| table | 主表 |
| sub_table | 正要新建的子表名 |
| start_date | 范围分界开始值 |
| end_date | 范围分界结束值 |
#!/usr/bin/python3
import psycopg2
from config import config
#example: create table tbl_game_android_step_log_2021_07 PARTITION OF tbl_game_android_step_log FOR VALUES FROM ('2021-07-01') TO ('2021-08-01');
def create_table(db, table, sub_table, start_date, end_date):
""" create subtable in the PostgreSQL database"""
command = "create table {0} PARTITION OF {1} FOR VALUES FROM ('{2[0]}') TO ('{2[1]}');".format(sub_table, table, (start_date, end_date))
conn = None
try:
# read the connection parameters
params = config(section = db)
# connect to the PostgreSQL server
conn = psycopg2.connect(**params)
cur = conn.cursor()
# create table one by one
cur.execute(command)
# close communication with the PostgreSQL database server
cur.close()
# commit the changes
conn.commit()
except (Exception, psycopg2.DatabaseError) as error:
print(error)
finally:
if conn is not None:
conn.close()
main.py：主文件；通过执行main生成分区表。
示例：
#!/usr/bin/python3
import datetime
from datetime import date
from dateutil.relativedelta import *
from pg_add_partition_table import create_table
#Get the 1st day of the next month
def get_next_month_first_day(d):
return date(d.year + (d.month == 12), d.month == 12 or d.month + 1 , 1)
def create_sub_table(db, table):
# Get current date
d1 = date.today()
# Get next month's date
d2 = d1 + relativedelta(months=+1)
# Get the 1st day of the next month;As the starting value of the partitioned table
start_date = get_next_month_first_day(d1)
# Gets the 1st of the next two months as the end value of the partitioned table
end_date = get_next_month_first_day(d2)
# get sub table name
getmonth = datetime.datetime.strftime(d2, '%Y_%m')
sub_table = table + '_' + getmonth
create_table(db, table, sub_table, start_date, end_date)
if __name__ == '__main__':
create_sub_table('test', 'tbl_game_android_step_log');
上面示例单独为表tbl_game_android_step_log；创建分区；若多个表；用for语句处理
# 多表操作
for table in ['tbl_game_android_step_log', 'tbl_game_android_game_log','tbl_game_android_pay_log']:
create_sub_table('test', table);
演示之前：
adsas=> select * from pg_partition_tree('tbl_game_android_step_log');
relid | parentrelid | isleaf | level
-----------------------------------+---------------------------+--------+-------
tbl_game_android_step_log | | f | 0
tbl_game_android_step_log_2020_12 | tbl_game_android_step_log | t | 1
(2 rows)
演示之后：
adsas=> select * from pg_partition_tree('tbl_game_android_step_log');
relid | parentrelid | isleaf | level
-----------------------------------+---------------------------+--------+-------
tbl_game_android_step_log | | f | 0
tbl_game_android_step_log_2020_12 | tbl_game_android_step_log | t | 1
tbl_game_android_step_log_2021_01 | tbl_game_android_step_log | t | 1
Partition key: RANGE (visit_time)
Partitions: tbl_game_android_step_log_2020_12 FOR VALUES FROM ('2020-12-01 00:00:00') TO ('2021-01-01 00:00:00'),
tbl_game_android_step_log_2021_01 FOR VALUES FROM ('2021-01-01 00:00:00') TO ('2021-02-01 00:00:00')
