PythonSqlite3:如何使用键盘中断等快速、干净地中断长时间运行的查询
使用Python中的模块,当接收到SIGINT(例如Control-C)时,长时间运行的查询不会很快中断/取消。sqlite3提供了一个方法,但是没有关于如何使用它的示例 有没有一种简单的方法可以中断/取消通过Python/sqlite3运行的长时间查询? 为了进行说明,首先生成一个测试数据库和表:PythonSqlite3:如何使用键盘中断等快速、干净地中断长时间运行的查询,sqlite,python,Sqlite,Python,使用Python中的模块,当接收到SIGINT(例如Control-C)时,长时间运行的查询不会很快中断/取消。sqlite3提供了一个方法,但是没有关于如何使用它的示例 有没有一种简单的方法可以中断/取消通过Python/sqlite3运行的长时间查询? 为了进行说明,首先生成一个测试数据库和表: import sqlite3 from random import randint conn = sqlite3.connect("randtable.db", 10.0) cursor = co
import sqlite3
from random import randint
conn = sqlite3.connect("randtable.db", 10.0)
cursor = conn.cursor()
cursor.execute("CREATE TABLE randint (id integer, rand integer)")
for i in range(1000000):
if i % 1000 == 0:
print ("{0}...".format(i))
rand = randint(0,1000000)
cursor.execute("INSERT INTO randint VALUES ({0},{1})".format(i,rand))
conn.commit()
conn.close()
然后在终端中执行长时间运行的Python/sqlite3脚本,并尝试使用Control-C中断该脚本:
from __future__ import print_function
import sqlite3
def main():
# Long running query (pathological by design)
statement ='''
SELECT DISTINCT a.id,a.rand
FROM randint a
JOIN randint b ON a.id=b.rand
JOIN randint c ON a.id=c.rand
JOIN randint d ON a.id=d.rand
JOIN randint e ON a.id=e.rand
JOIN randint f ON a.id=f.rand
JOIN randint g ON a.id=g.rand
JOIN randint h ON a.id=h.rand
ORDER BY a.id limit 10'''
conn = sqlite3.connect('randtable.sqlite', 10.0)
cursor = conn.cursor()
print ("Executing query")
cursor.execute(statement)
rows = cursor.fetchall()
print ("ROWS:")
for row in rows:
print (" ", row)
conn.close()
return
if __name__ == "__main__":
main()
在终端中运行上述脚本,然后按Control-C(或以其他方式发送SIGINT)将最终取消查询和脚本,但这可能需要相当长的时间,也可能需要很多分钟。当按下Control-C时,在中运行的完全相同的查询几乎立即被取消
提前谢谢 回答我自己的问题,因为我想我已经解决了。下面是我想到的,对这段代码的任何评论都将不胜感激
#!/usr/bin/env python
from __future__ import print_function
import sqlite3
import threading
import signal
import os
import time
conn = None
shutdown = False
def main():
global conn
# Long running query (pathological by design)
statement ='''
SELECT DISTINCT a.id,a.rand
FROM randint a
JOIN randint b ON a.id=b.rand
JOIN randint c ON a.id=c.rand
JOIN randint d ON a.id=d.rand
JOIN randint e ON a.id=e.rand
JOIN randint f ON a.id=f.rand
JOIN randint g ON a.id=g.rand
JOIN randint h ON a.id=h.rand
ORDER BY a.id limit 10'''
conn = sqlite3.connect('randtable.sqlite', 10.0)
cursor = conn.cursor()
print ("Executing query")
try:
cursor.execute(statement)
except Exception as err:
if str(err) != "interrupted":
print ("Database error: {0}".format(str(err)))
return None
rows = cursor.fetchall()
print ("ROWS:")
for row in rows:
print (" ", row)
conn.close()
conn = None
return
def interrupt(signum, frame):
global conn
global shutdown
print ("Interrupt requested")
if conn:
conn.interrupt()
if __name__ == "__main__":
signal.signal(signal.SIGINT, interrupt)
mainthread = threading.Thread(target=main)
mainthread.start()
while mainthread.isAlive():
time.sleep(0.2)
你的答案涵盖了这一点,但是(昨天我忘记了——对不起!)我记得我答应过要写一个答案,所以这里有另一个版本证明你可以不用全局变量来完成这一点。我在这里还使用了
线程化.Event
而不是信号来演示有几种不同的方法可以向线程发出信号,表示是时候做点什么了(但出于您的目的,请坚持使用信号,因为这非常适合对Ctrl+C作出反应):
那么呢?@DaveJones是的,正如问题最上面提到的,这应该是解决问题的方法。但是如何使用真正的代码呢?我尝试过但没有成功,也许我在Python方面的专业知识不足让我看不到显而易见的东西。对不起,错过了这一点-我现在必须去接我的女儿,但我会在一点时间内写一个例子(它将类似于
threading.Thread(target=conn.interrupt).start()
,如果这足以让您同时开始学习的话)。
import sqlite3
import time
import threading
# Background thread that'll kill our long running query after 1 second
def kill_it(connection, event):
event.wait()
time.sleep(1)
connection.interrupt()
# Make some tables with lots of data so we can make a long running query
def big_query(conn, kill_event):
print('Making big tables')
conn.execute(
"CREATE TABLE foo (i integer primary key, s text);")
conn.execute(
"CREATE TABLE bar (j integer primary key, s text);")
conn.execute(
"INSERT INTO foo VALUES %s" % ", ".join("(%d, 'foo')" % i for i in range(10000)))
conn.execute(
"INSERT INTO bar VALUES %s" % ", ".join("(%d, 'bar')" % i for i in range(10000)))
kill_event.set()
print('Running query')
cur = conn.cursor()
cur.execute(
"SELECT * FROM foo, bar")
print(len(cur.fetchall()))
def main():
conn = sqlite3.connect('foo.db')
kill_event = threading.Event()
kill_thread = threading.Thread(target=kill_it, args=(conn, kill_event))
kill_thread.start()
big_query(conn, kill_event)
kill_thread.join()
if __name__ == '__main__':
main()