Can I use pymysql.connect() with “with” statement?

我的梦境 提交于 2020-08-24 05:39:51

问题


The following is listed as example in pymysql:

conn = pymysql.connect(...)
with conn.cursor() as cursor:
    cursor.execute(...)
    ...
conn.close()

Can I use the following instead, or will this leave a lingering connection? (it executes successfully)

import pymysql
with pymysql.connect(...) as cursor:
    cursor.execute('show tables')

(python 3, latest pymysql)


回答1:


This does not look safe, if you look here, the __enter__ and __exit__ functions are what are called in a with clause. For the pymysql connection they look like this:

def __enter__(self):
    """Context manager that returns a Cursor"""
    return self.cursor()

def __exit__(self, exc, value, traceback):
    """On successful exit, commit. On exception, rollback"""
    if exc:
        self.rollback()
    else:
        self.commit()

So it doesn't look like the exit clause closes the connection, which means it would be lingering. I'm not sure why they did it this way. You could make your own wrappers that do this though.

You could recycle a connection by creating multiple cursors with it (the source for cursors is here) the cursor methods look like this:

def __enter__(self):
    return self

def __exit__(self, *exc_info):
    del exc_info
    self.close()

So they do close themselves. You could create a single connection and reuse it with multiple cursors in with clauses.

If you want to hide the logic of closing connections behind a with clause, e.g. a context manager, a simple way to do it would be like this:

from contextlib import contextmanager
import pymysql


@contextmanager
def get_connection(*args, **kwargs):
    connection = pymysql.connect(*args, **kwargs)
    try:
        yield connection
    finally:
        connection.close()

You could then use that context manager like this:

with get_connection(...) as con:
    with con.cursor() as cursor:
        cursor.execute(...)



回答2:


As it was pointed out, the Cursor takes care of itself, but all the Connection's support for context manager was removed completely just a few days ago, so the only option now is to write yours:

https://github.com/PyMySQL/PyMySQL/pull/763

https://github.com/PyMySQL/PyMySQL/issues/446




回答3:


As an alternative to this, since I wanted to support the context manager pattern for a connection, I implemented it with a monkey patch. Not the best approach, but it's something.

import pymysql


MONKEYPATCH_PYMYSQL_CONNECTION = True


def monkeypatch_pymysql_connection():
    Connection = pymysql.connections.Connection

    def enter_patch(self):
        return self

    def exit_patch(self, exc, value, traceback):
        try:
            self.rollback()  # Implicit rollback when connection closed per PEP-249
        finally:
            self.close()

    Connection.__enter__ = enter_patch
    Connection.__exit__ = exit_patch


if MONKEYPATCH_PYMYSQL_CONNECTION:
    monkeypatch_pymysql_connection()
    MONKEYPATCH_PYMYSQL_CONNECTION = False  # Prevent patching more than once

This approach worked for my use case. I would prefer to have __enter__ and __exit__ methods in the Connection class. That approach, however, was rejected by the developers when they addressed the issue in late 2018.



来源:https://stackoverflow.com/questions/31214658/can-i-use-pymysql-connect-with-with-statement

易学教程内所有资源均来自网络或用户发布的内容,如有违反法律规定的内容欢迎反馈
该文章没有解决你所遇到的问题?点击提问,说说你的问题,让更多的人一起探讨吧!