Hadoop MapReduce处理MovieLens ml-100k数据集

数据集介绍

待补充
user id | item id | rating | timestamp.

---

每个用户的平均评分

map阶段

#coding=utf-8
import sys

for line in sys.stdin:
    line = line.strip().split()
    #userid rating
    print "%s\t%s" % (line[0], line[2])

[root@DW1 code]# cat u.data|python rating_mapper.py
...
880	3
716	5
276	1
13	2
12	3

---

reduce阶段

#coding=utf-8
import sys

curr_userid = None
curr_rating = 0
count = 0

for line in sys.stdin:
    userid, rating = line.strip().split('\t')

    if not curr_userid:
        curr_userid = userid
    if userid != curr_userid:
        print curr_userid + "\t" + str(curr_rating/count)
        curr_rating = 0
        count = 0
        curr_userid = userid
    curr_rating += int(rating)
    count +=