◢ <實作七> | <回課程大綱> ▲ | <實作九> ◣

實作八 Lab8

練習豬的拉丁語
Pig Latin in Practice

• 首先使用您的帳號，連線至 hadoop.nchc.org.tw

Aggregation (Local Mode)

~$ wget http://hadoop.nchc.org.tw/excite-small.log
~$ pig -x local
grunt> log = LOAD 'excite-small.log' AS (user, timestamp, query);
grunt> grpd = GROUP log BY user;
grunt> cntd = FOREACH grpd GENERATE group, COUNT(log);
grunt> STORE cntd INTO 'lab8_out1';
grunt> quit
~$ head lab8_out1

Filter (Local Mode)

~$ pig -x local
grunt> log = LOAD 'excite-small.log' AS (user, timestamp, query);
grunt> grpd = GROUP log BY user;
grunt> cntd = FOREACH grpd GENERATE group, COUNT(log) AS cnt;
grunt> fltrd = FILTER cntd BY cnt > 50;
grunt> STORE fltrd INTO 'lab8_out2';
grunt> quit
~$ head lab8_out2

Sorting (Local Mode)

~$ pig -x local
grunt> log = LOAD 'excite-small.log' AS (user, timestamp, query);
grunt> grpd = GROUP log BY user;
grunt> cntd = FOREACH grpd GENERATE group, COUNT(log) AS cnt;
grunt> fltrd = FILTER cntd BY cnt > 50;
grunt> srtd = ORDER fltrd BY cnt;
grunt> STORE srtd INTO 'lab8_out3';
grunt> quit
~$ head lab8_out3

Connect Pig to Hadoop (Full Distributed Mode)

~$ hadoop fs -put excite-small.log .
~$ pig
grunt> log = LOAD 'excite-small.log' AS (user, timestamp, query);
grunt> grpd = GROUP log BY user;
grunt> cntd = FOREACH grpd GENERATE group, COUNT(log);
grunt> STORE cntd INTO 'lab8_out1';
grunt> quit
~$ hadoop fs -cat lab8_out1/part-00000