Hive中使用python

使用方法如下

set mapred.reduce.tasks = 300;

add file /home/work/process.py;

insert overwrite  directory '/mydir/'

select * 
from(
 from(
     select id, name
     from hive_table_one
     where name = '張三'
  )one
  join 
  (
     select id, name
     from hive_table_two
     where name = '李四'
  )two
  on one.id = two.id
  reduce one.id, one.name, two.id, two.name
  using '/home/sharelib/python/bin/python process.py'
  as id, name
)redall

在process.py腳本處理Hive表中的NULL數據時，需要注意一下。

# 判斷name是否為NULL的語句如下
if (name == '\N')

#如果是先查詢出結果，存成文本，再進行處理。那么就會是
if (name == 'NULL')