处理按第一列聚合第二列的代码示例

#coding=gbk
import os
import sys
import argparse
import commands
import logging


status, username = commands.getstatusoutput('whoami')
logging.basicConfig(level = logging.DEBUG,
        format=username + ' %(asctime)s %(filename)s[line:%(lineno)d] %(levelname)s %(message)s',
        datefmt = '%Y-%m-%d %H:%M:%S',
        filename = '.log' + __file__ ,
        filemode = 'a')

def perror_and_exit(message, status = -1):
    sys.stderr.write(message + '\n')
    logging.info(message + 'sys exit')
    sys.exit(status)

count = 0
def handle_keywords(userid, keywords, fw):
    global count
    count += 1
    pass

if __name__ == "__main__":
    directory = "/home/junfeng"
    if len(sys.argv) != 2:
        perror_and_exit("Usage:python %s keyword or idea" % (__file__ + ""))
    if sys.argv[1] == "keyword":
        filename = os.path.join(directory, "userid_showword.active.tradeid5401.20130329.txt")
        print filename
        fw = file("userid_showwords_brand.txt", "w")
        last_userid = None
        keywords = []
        for line in file(filename):
            line = line.strip().split()
            userid = line[0]
            keyword = line[1]

            if userid == last_userid:
                keywords.append(keyword)
            else:
                if len(keywords) > 0:
                    handle_keywords(last_userid, keywords, fw)
                    pass
                last_userid = userid
                keywords = [keyword]
        handle_keywords(last_userid, keywords, fw)
        print keywords
        global count
        print count
        pass
    elif sys.argv[1] == "idea":
        filename = os.path.join(directory, "userid_title_desc1_desc2.active.tradeid5401.20130329.txt")
        for line in file(filename):
            line = line.strip().split()
        pass