psycopg2 postgresql driver for python don't support prepared statement Direct-阿里云开发者社区

psycopg2 postgresql driver for python don't support prepared statement Direct

2016-03-30 3472

版权

本文内容由阿里云实名注册用户自发贡献，版权归原作者所有，阿里云开发者社区不拥有其著作权，亦不承担相应法律责任。具体规则请查看《阿里云开发者社区用户服务协议》和《阿里云开发者社区知识产权保护指引》。如果您发现本社区中有涉嫌抄袭的内容，填写侵权投诉表单进行举报，一经查实，本社区将立刻删除涉嫌侵权内容。

本文涉及的产品

云原生数据库 PolarDB MySQL 版，Serverless 5000PCU 100GB

云原生数据库 PolarDB PostgreSQL 版，企业版 4核16GB

云原生数据库 PolarDB MySQL 版，通用型 2核4GB 50GB

简介：

前面使用py-postgresql测试过PostgreSQL性能, 可能是这个驱动效率较低, 我们接下来使用psycopg2测试一下.

psycopg2使用libpq接口, 支持2PC, 支持异步提交等, 但是不支持绑定变量.

安装

[root@localhost ~]# . /home/postgres/.bash_profile 
root@localhost-> which pg_config
/opt/pgsql9.3.5/bin/pg_config
[root@localhost ~]# pip3.4 install psycopg2 --upgrade
或
[root@localhost ~]# pip3.4 install psycopg2

异步操作, 对数据库端来说意义不大, 因为在执行的话, 不能再次提交SQL请求.

如下 :

import psycopg2
import time
import select
conn = psycopg2.connect(database="postgres", user="postgres", password="postgres", host="/data01/pgdata/pg_root", port="1921", async=True)
>>> def wait(conn):
...     while 1:
...         state = conn.poll()
...         if state == psycopg2.extensions.POLL_OK:
...             break
...         elif state == psycopg2.extensions.POLL_WRITE:
...             select.select([], [conn.fileno()], [])
...         elif state == psycopg2.extensions.POLL_READ:
...             select.select([conn.fileno()], [], [])
...         else:
...             raise psycopg2.OperationalError("poll() returned %s" % state)
... 

wait(conn)
curs = conn.cursor()
curs.execute("insert into tt values(%(id)s, 'digoal.zhou', 32, 'digoal@126.com', '276732431')", {"id": 1})
curs.execute("insert into tt values(%(id)s, 'digoal.zhou', 32, 'digoal@126.com', '276732431')", {"id": 1})
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
psycopg2.ProgrammingError: execute cannot be used while an asynchronous query is underway
>>> wait(curs.connection)
>>> wait(curs.connection)
>>> curs.execute("insert into tt values(%(id)s, 'digoal.zhou', 32, 'digoal@126.com', '276732431')", {"id": 1})
>>> wait(curs.connection)
>>> curs.execute("insert into tt values(%(id)s, 'digoal.zhou', 32, 'digoal@126.com', '276732431')", {"id": 1})
>>> wait(curs.connection)
>>> conn.poll()
0
>>> curs.execute("insert into tt values(%(id)s, 'digoal.zhou', 32, 'digoal@126.com', '276732431')", {"id": 1})
>>> conn.poll()
0
>>> curs.execute("insert into tt values(%(id)s, 'digoal.zhou', 32, 'digoal@126.com', '276732431')", {"id": 1})
>>> conn.poll()
0
>>> psycopg2.extensions.POLL_OK
0
>>> psycopg2.extensions.POLL_WRITE
2
>>> psycopg2.extensions.POLL_READ
1

同时psycopg2不支持prepared statement

>>> curs.prepare("insert into tt values(%(id)s, 'digoal.zhou', 32, 'digoal@126.com', '276732431')")
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: 'psycopg2._psycopg.cursor' object has no attribute 'prepare'
>>> dir(curs)
['__class__', '__delattr__', '__dir__', '__doc__', '__enter__', '__eq__', '__exit__', '__format__', '__ge__', '__getattribute__', '__gt__', '__hash__', '__init__', '__iter__', '__le__', '__lt__', '__ne__', '__new__', '__next__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr__', '__sizeof__', '__str__', '__subclasshook__', 'arraysize', 'binary_types', 'callproc', 'cast', 'close', 'closed', 'connection', 'copy_expert', 'copy_from', 'copy_to', 'description', 'execute', 'executemany', 'fetchall', 'fetchmany', 'fetchone', 'itersize', 'lastrowid', 'mogrify', 'name', 'nextset', 'query', 'row_factory', 'rowcount', 'rownumber', 'scroll', 'scrollable', 'setinputsizes', 'setoutputsize', 'statusmessage', 'string_types', 'typecaster', 'tzinfo_factory', 'withhold']

为什么这么说呢?, 调用 curs.execute("insert into tt values(%(id)s, 'digoal.zhou', 32, 'digoal@126.com', '276732431')", {"id": 1})

你觉得是绑定变量吗? 看看PostgreSQL的日志吧 :

2015-02-05 22:37:32.901 CST,"postgres","postgres",69785,"[local]",54d37ee8.11099,1,"idle",2015-02-05 22:32:08 CST,3/1996971,0,LOG,00000,"statement: insert into tt values(1, 'digoal.zhou', 32, 'digoal@126.com', '276732431')",,,,,,,,"exec_simple_query, postgres.c:890",""

调用的是exec_simple_query接口, 当然不是绑定变量.

间接的使用绑定变量, 但实际上SQL调用的还是 exec_simple_query接口, 只是在处理execute这个SQL时使用到了绑定变量.

这个是较老的用法.

>>> curs.execute("prepare pre1(int) as insert into tt values($1, 'digoal.zhou', 32, 'digoal@126.com', '276732431')")
>>> conn.poll()
0
>>> curs.execute("execute pre1(1)")
>>> conn.poll()
0
>>> curs.execute("execute pre1(1)")
>>> conn.poll()
0

PostgreSQL日志 :

2015-02-05 22:50:42.678 CST,"postgres","postgres",69785,"[local]",54d37ee8.11099,5,"idle",2015-02-05 22:32:08 CST,3/1996975,0,LOG,00000,"statement: prepare pre1(int) as insert into tt values($1, 'digoal.zhou', 32, 'digoal@126.com', '276732431')",,,,,,,,"exec_simple_query, postgres.c:890",""
2015-02-05 22:50:59.425 CST,"postgres","postgres",69785,"[local]",54d37ee8.11099,6,"idle",2015-02-05 22:32:08 CST,3/1996976,0,LOG,00000,"statement: execute pre1(1)","prepare: prepare pre1(int) as insert into tt values($1, 'digoal.zhou', 32, 'digoal@126.com', '276732431')",,,,,,,"exec_simple_query, postgres.c:890",""
2015-02-05 22:51:22.425 CST,"postgres","postgres",69785,"[local]",54d37ee8.11099,7,"idle",2015-02-05 22:32:08 CST,3/1996977,0,LOG,00000,"statement: execute pre1(1)","prepare: prepare pre1(int) as insert into tt values($1, 'digoal.zhou', 32, 'digoal@126.com', '276732431')",,,,,,,"exec_simple_query, postgres.c:890",""

效率如何呢?

使用8个线程插入100W数据, 耗时41秒.

py-postgresql驱动耗时226秒, psycopg2效率高了很多, 接近pgbench的16秒了.

[root@localhost ~]# cat t.py
import psycopg2
import time
import threading



class n_t(threading.Thread):   #The timer class is derived from the class threading.Thread
  def __init__(self, num):
    threading.Thread.__init__(self)
    self.thread_num = num

  def run(self): #Overwrite run() method, put what you want the thread do here
    conn = psycopg2.connect(database="postgres", user="postgres", password="postgres", host="/data01/pgdata/pg_root", port="1921")
    curs = conn.cursor()
    conn.autocommit=True

    start_t = time.time()
    print("TID:" + str(self.thread_num) + " " + str(start_t))

    for i in range(self.thread_num*125000, (self.thread_num+1)*125000):
      curs.execute("insert into tt values(%(id)s, 'digoal.zhou', 32, 'digoal@126.com', '276732431')", {"id": i})

    stop_t = time.time()
    print("TID:" + str(self.thread_num) + " " + str(stop_t))
    print(stop_t-start_t)

def test():
  t_names = dict()
  for i in range(0,8):
    t_names[i] = n_t(i) 
    t_names[i].start()
  return

if __name__ == '__main__':
  test()

测试结果 :

[root@localhost ~]# python t.py
TID:0 1423148404.5820847
TID:1 1423148404.5828164
TID:2 1423148404.5850544
TID:3 1423148404.58532
TID:4 1423148404.5861893
TID:5 1423148404.5867805
TID:6 1423148404.5882406
TID:7 1423148404.588671
TID:2 1423148445.146449
40.561394691467285
TID:7 1423148445.2051861
40.616515159606934
TID:6 1423148445.2321012
40.64386057853699
TID:1 1423148445.263236
40.68041968345642
TID:5 1423148445.2703242
40.68354368209839
TID:0 1423148445.2967775
40.71469283103943
TID:3 1423148445.3065617
40.72124171257019
TID:4 1423148445.3345916
40.74840235710144

postgres=# select count(*),count(distinct id) from tt;
  count  |  count  
---------+---------
 1000000 | 1000000
(1 row)

[参考]

1. https://wiki.postgresql.org/wiki/Python

2. https://wiki.postgresql.org/wiki/Using_psycopg2_with_PostgreSQL

3. http://initd.org/psycopg/docs/index.html