Pipe in PySpark

官网的文档声明，pipe管道将pyspark的RDD对象fork给外部进程，然后再返回RDD对象

>>> sc.parallelize(['1', '2', '', '3']).pipe('cat').collect()
[u'1', u'2', u'', u'3']

---

python代码如下，外部进程为cat命令

#!/usr/bin/python
# -*- coding: utf-8 -*-
import sys
import os
from pyspark import SparkContext
from pyspark import SparkConf

if __name__ == '__main__':
    sc = SparkContext()
    data  = ["hi","hello","how","are","you"]
    dataRDD = sc.parallelize(data)
    pipeRDD = dataRDD.pipe('cat').collect()
    print pipeRDD

---

再提交任务即测试成功

spark-submit –master=yarn-client xx.py