python pipe stdout 实现cat|grep 功能

从hdfs里获取希望的数据：

import subprocess

for day in range(22, 23):
    for h in range(17, 24):
        filename = "metadata-2018-10-%02d-%02d.txt" % (day, h)
        cmd = "hdfs dfs -text /flume/metadata/2018/10/%02d/%02d/*.snappy" % (day, h)
        print(cmd)
        #cmd = "cat *.py"
        cmd = subprocess.Popen(cmd, shell=True, stdout=subprocess.PIPE)
        f = open(filename, "w")
        for line in cmd.stdout:
            try:
              arr = line.split("^")
              if len(line) > 100 and arr[6] == "6":
                #print(line)
                f.write(" ".join(arr[:32]) + " " + arr[95] + "\n")
            except Exception as e:
                print(e, "fuck error", line)
        f.close()
        #import sys
        #sys.exit(0)