Nov
30
最近项目上有需要，大概就是有一个list里的东西需要处理，例如一堆文件什么的，于是有一个file_processor——按顺序处理一个文件列表。简单封装了一下multiprocessing这个库，发现用起来很方便，很轻松地就实现了多进程并行处理（进程间无交互）：
multiprocess.py 则是这样的：
import multiprocess
slices = multiprocess.split_list(filelist, 8) #分成8份
processes = map(lambda slice: multiprocess.spawn(file_processor, slice), slices)
sys.exit(multiprocess.start_and_join(processes))
slices = multiprocess.split_list(filelist, 8) #分成8份
processes = map(lambda slice: multiprocess.spawn(file_processor, slice), slices)
sys.exit(multiprocess.start_and_join(processes))
multiprocess.py 则是这样的：
#!/usr/bin/python
#coding: utf-8
import sys
from multiprocessing import Process
def split_list(data, n_slice, hash_func=lambda i, d: i): #default: sequential
slices = []
for i in range(n_slice):
slices.append([])
for i, d in enumerate(data):
slices[hash_func(i, d) % n_slice].append(d)
return slices
def spawn(target, *args, **kwargs):
return Process(target=target, args=args, kwargs=kwargs)
def start_and_join(processes, killall_if_fail=True):
for p in processes:
p.start()
exitcode = 0
for p in processes:
p.join()
if p.exitcode != 0:
exitcode = p.exitcode
break
if exitcode != 0:
for p in processes:
if killall_if_fail and p.is_alive():
p.terminate()
return exitcode
#coding: utf-8
import sys
from multiprocessing import Process
def split_list(data, n_slice, hash_func=lambda i, d: i): #default: sequential
slices = []
for i in range(n_slice):
slices.append([])
for i, d in enumerate(data):
slices[hash_func(i, d) % n_slice].append(d)
return slices
def spawn(target, *args, **kwargs):
return Process(target=target, args=args, kwargs=kwargs)
def start_and_join(processes, killall_if_fail=True):
for p in processes:
p.start()
exitcode = 0
for p in processes:
p.join()
if p.exitcode != 0:
exitcode = p.exitcode
break
if exitcode != 0:
for p in processes:
if killall_if_fail and p.is_alive():
p.terminate()
return exitcode
