poorman

하둡 어플리케이션은 보통 Mapper -> Shuffle -> Reducer 순으로 작업을 진행한다.

파이썬 예제와 함께 각각의 결과물을 확인한다.

1. Mapper

#!/usr/bin/env python import sys for line in sys.stdin: line = line.strip() keys = line.split() for key in keys: value = 1 print( "%s\t%d" % (key, value) )

$ cat wordcount_mapper.py | python ./wordcount_mapper.py > output_mapper.txt
$ cat output_mapper.txt
import	1
sys	1
for	1
line	1
in	1
sys.stdin:	1
line	1
=	1
line.strip()	1
keys	1
=	1
line.split()	1
for	1
key	1
in	1
keys:	1
value	1
=	1
1	1
print("{0}\t{1}".format(key,value))	1

2. Shuffle

$ cat output_mapper.txt | sort > output_sort.txt
$ cat output_sort.txt
=	1
=	1
=	1
1	1
for	1
for	1
import	1
in	1
in	1
key	1
keys:	1
keys	1
line	1
line	1
line.split()	1
line.strip()	1
print("{0}\t{1}".format(key,value))	1
sys	1
sys.stdin:	1
value	1

3. Reducer

#!/usr/bin/env python import sys last_key = None running_total = 0 for input_line in sys.stdin: input_line = input_line.strip() this_key, value = input_line.split("\t", 1) value = int(value) if last_key == this_key: running_total += value else: if last_key: print( "%s\t%d" % (last_key, running_total) ) running_total = value last_key = this_key if last_key == this_key: print( "%s\t%d" % (last_key, running_total) )

$ cat output_sort.txt | python wordcount_reducer.py > output_reducer.txt
$ cat output_reducer.txt 
=	3
1	1
for	2
import	1
in	2
key	1
keys:	1
keys	1
line	2
line.split()	1
line.strip()	1
print("{0}\t{1}".format(key,value))	1
sys	1
sys.stdin:	1
value	1

4. Mapper | Shuffle | Reducer

$ cat wordcount_mapper.py | python wordcount_mapper.py | sort | python wordcount_reducer.py > output.txt
$ cat output.txt
=	3
1	1
for	2
import	1
in	2
key	1
keys:	1
keys	1
line	2
line.split()	1
line.strip()	1
print("{0}\t{1}".format(key,value))	1
sys	1
sys.stdin:	1
value	1

딥 러닝에 대해 독학을 하면서 정리한 걸 적고 있습니다.

전공과 무관하며 전문적인 지식이 아니므로 개인적인 의견과 부족하고 틀린 점이 많습니다.

추가 지식 및 잘못된 점을 지적해주시면 공부하는데 많은 도움이 되겠습니다. 감사합니다^^

- 푸어맨

[Reference]

(Writing Hadoop Applications in Python with Hadoop Streaming) http://www.glennklockwood.com/data-intensive/hadoop/streaming.htm

(하둡 스트리밍을 활용한 word count 예제) http://blog.acronym.co.kr/606

(파이썬 문자열 관련함수) http://agiantmind.tistory.com/31