線程中的subprocess.Popen

Question

我有很多文件（超過4000個），我想同時加載到PostgreSQL中。 我已將它們分成4個不同的文件列表，我想要一個線程迭代加載數據的每個列表。

我遇到的問題是我使用os.system來調用加載程序但這會阻止其他線程同時運行。 如果我使用subprocess.Popen然后它們同時運行但是線程認為它們已經完成了execeuting，所以移動到我的腳本的下一部分。

我這樣做是對的嗎？ 或者是否有更好的方法從線程內調用子進程。

def thread1Load(self, thread1fileList):
    connectionstring = settings.connectionstring
    postgreshost = settings.postgreshost
    postgresdatabase = settings.postgresdatabase
    postgresport = settings.postgresport
    postgresusername = settings.postgresusername
    postgrespassword = settings.postgrespassword

    tablename = None
    encoding = None
    connection = psycopg2.connect(connectionstring)

    for filename in thread1fileList:
        load_cmd = #load command
        run = subprocess.Popen(load_cmd, shell=True)
    print "finished loading thread 1"


def thread2Load(self, thread2fileList):
    connectionstring = settings.connectionstring
    postgreshost = settings.postgreshost
    postgresdatabase = settings.postgresdatabase
    postgresport = settings.postgresport
    postgresusername = settings.postgresusername
    postgrespassword = settings.postgrespassword

    tablename = None

    connection = psycopg2.connect(connectionstring)
    for filename in thread2fileList:
        load_cmd = #load command            
        run = subprocess.Popen(load_cmd, shell=True)
    print "finished loading thread 2"


def thread3Load(self, thread3fileList):
    connectionstring = settings.connectionstring
    postgreshost = settings.postgreshost
    postgresdatabase = settings.postgresdatabase
    postgresport = settings.postgresport
    postgresusername = settings.postgresusername
    postgrespassword = settings.postgrespassword

    tablename = None
    connection = psycopg2.connect(connectionstring)

    for shapefilename in thread3fileList:
        load_cmd = #load command
        run = subprocess.Popen(load_cmd, shell=True)
    print "finished loading thread 3"

def thread4Load(self, thread4fileList):
    connectionstring = settings.connectionstring
    postgreshost = settings.postgreshost
    postgresdatabase = settings.postgresdatabase
    postgresport = settings.postgresport
    postgresusername = settings.postgresusername
    postgrespassword = settings.postgrespassword

    tablename = None

    connection = psycopg2.connect(connectionstring)

    for filename in thread4fileList:
        load_cmd = #load command
        run = subprocess.Popen(load_cmd, shell=True)

    print "finished loading thread 4"


def finishUp(self):
    print 'finishing up'


def main():
load = Loader()

thread1 = threading.Thread(target=(load.thread1Load), args=(thread1fileList, ))
thread2 = threading.Thread(target=(load.thread2Load), args=(thread2fileList, ))
thread3 = threading.Thread(target=(load.thread3Load), args=(thread3fileList, ))
thread4 = threading.Thread(target=(load.thread4Load), args=(thread4fileList, ))
threads = [thread1, thread2, thread3, thread4]
for thread in threads:
    thread.start()
    thread.join()


load.finishUp(connectionstring)

if __name__ == '__main__':
main()

Answer 1

• 不要重復自己 。 一個threadLoad方法就足夠了。 這樣，如果您需要修改方法中的某些內容，則無需在4個不同的位置進行相同的修改。
• 使用run.communicate()來阻止，直到子run.communicate()完成。

• 這將啟動一個線程，然后阻塞直到該線程完成，然后啟動另一個線程等：

   for thread in threads: thread.start() thread.join() 

  相反，首先啟動所有線程，然后加入所有線程：

   for thread in threads: thread.start() for thread in threads: thread.join() 

---

import subprocess
import threading


class Loader(object):
    def threadLoad(self, threadfileList):
        connectionstring = settings.connectionstring
        ...
        connection = psycopg2.connect(connectionstring)

        for filename in threadfileList:
            load_cmd =  # load command
            run = subprocess.Popen(load_cmd, shell=True)
            # block until subprocess is done
            run.communicate()
        name = threading.current_thread().name
        print "finished loading {n}".format(n=name)

    def finishUp(self):
        print 'finishing up'


def main():
    load = Loader()
    threads = [threading.Thread(target=load.threadLoad, args=(fileList, ))
               for fileList in (thread1fileList, thread2fileList,
                                thread3fileList, thread4fileList)]
    for thread in threads:
        thread.start()
    for thread in threads:
        thread.join()

    load.finishUp(connectionstring)

if __name__ == '__main__':
    main()