hadoop-pig学习笔记

A1 = LOAD '/luo/lzttxt01.txt' AS (col1:chararray,col2:int,col3:int,col4:int,col5:double,col6:double);
B1 = GROUP A1 BY (col2,col3,col4);
C1 = FOREACH B1 GENERATE FLATTEN(group),AVG(A1.col5),AVG(A1.col6); ---这里的A1指的是B1里的A1，B1中有若干个A1;
STORE C1 INTO '/output1';

A = LOAD '/luo/txt.txt' AS (col1:chararray,col2:int,col3:int,col4:int,col5:double,col6:double);
B = GROUP A ALL;
C = FOREACH B GENERATE COUNT(A.col2);
DUMP C;

A1 = LOAD '/lu/lzttxt01.txt' AS (col1:chararray,col2:int,col3:int,col4:int,col5:double,col6:double);
B1 = GROUP A1 BY (col2,col3,col4);
C1 = FOREACH B1 {D = DISTINCT A1.col6; GENERATE group ,COUNT(D);};
DUMP C1;

A = LOAD '/lu1/b01.txt' AS (col1:int,col2:int,col3:int,col4:chararray,col5:chararray);
B = STREAM A THROUGH `awk '{if($4=="=") print $1"\t"$2"\t"$3"\t9999\t"$5;else print $0}'`; -- STREAM .. THROUGH ..调用shell语句 当第四列为“=”号时，将其替换为9999，否则就按照原样输出这一行
DUMP B

若是aa.pig 方式的话文件要放在root本地：pig -x mapreduce /luo/aa.pig 执行 或是 pig -x local test.pig

A = LOAD '/lu1/a.txt' AS (acol1:chararray,acol2:int,acol3:int);
B = LOAD '/lu1/c.txt' AS (bol1:int,bol2:chararray,bol3:int);
C = COGROUP A BY acol1,B BY bol2; ---cogroup 可以按多个关系中的字段进行分组
DUMP C;

A = LOAD '/lzt02/aa01.txt' AS (a:int,b:int);
B = LOAD '/lzt02/aa02.txt' AS (c:int,d:int);
C = UNION A,B;
D = GROUP C BY $0;按第一列进行分组
E = FOREACH D GENERATE FLATTEN(group),SUM(C.$1); $1第二列
DUMP E;

A1 = LOAD '/lzt02/aa03.txt' AS (a11:int,b11:chararray); 符合"*//*.qq.com/*"
B1 = FILTER A1 BY b11 matches '.*//.*\\.qq\\.com/.*';
C1 = FILTER B1 BY (a11 matches('\\d'))
DUMP B1;
.表示任意字符， * 表示任意次数，\.是对.的转义 ，/就是表示/这个字符；注意在引号中\\.才是和正则中的\.的一致
正则中\d表示匹配数字，在引号中必须用'\\d'

1.定义的数据结构为元组
A = LOAD 'luo.txt' AS (T : tuple(col1:int,col2:int,col3:int,col4:chararray,col5:chararray)) --使用与数据是这种格式的（1,2,3，5,2,4）

2.
STORE A into '$output_dir' --使用参数
pig -param output_dir="/home/my_outdir" my_pig_script.pig
3.load 多个目录下的数据：
A = LOAD '/abc/201{0,1}'
4.两个整数相除，想得到整数和浮点数
整数：(float)(col1/col2) 浮点数： (float)col1/col2
5.substring
B = FOREACH A GENERATE SUBSTRING(date，0,4)
6.拼接concat
A = LOAD '1.txt' AS (col1:chararray,col2:int);
B = FOREACH A GENERATE CONCAT(col1,(chararray)col2) ----多个字段进行拼接时使用concat嵌套：concat(a,concat(b,c))
7.join的用法,求两个数据表中重合的个数
A = LOAD '/lzt02/aa01.txt' AS (a:int,b:int);
B = LOAD '/lzt02/aa02.txt' AS (c:int,d:int);
C = JOIN A BY a,B BY d;
D = DISTINCT C;
E = GROUP D ALL;
F = FOREACH E GENERATE COUNT(D)

8.使用三目运算符“?:”
B = FOREACH A GENERATE col1,((col2 is null)?-1 :col2),col3

或
A = LOAD '1.txt' AS (a:int,b:tuple(x:int,y:int)); ----适用于2,(3,5)这样的数据
B = FOREACH A GENERATE a,FLATEEN(b);
C = FOREACH B GENERATE group ,SUM(B.x) AS S;
D = FOREACH C GENERATE group，(s is null)?-1 :s

9. 在第一列的每种组合下，第二列为3和6的数据分别有多少条
A = LOAD '/lzt02/aa01.txt' AS (a:int,b:int);
B = GROUP A BY a;
C = FOREACH B {
D = FILTER A BY b==3; ##这里的A是B中的A
E = FILTER A BY b==6;
GENERATE group,count(D),count(E);
}
DUMP C;B

10. A = LOAD '/lzt02/aa01.txt' AS (a:int,b:int);
B = LOAD '/lzt02/aa02.txt' AS (c:int,d:int);
C = JOIN A BY a LEFT OUTER,B BY d; ----c中有两张表的全部字段 ，遵循left join 原则
D = DISTINCT C;
E = GROUP D ALL;
F = FOREACH E GENERATE COUNT(D)

11.A表中有，但是B表中没有的数据
A = LOAD '/lzt02/aa01.txt' AS (a1:int,b1:int);
B = LOAD '/lzt02/aa02.txt' AS (a1:int,b1:int);
C = JOIN A BY a1 left outer,B BY a1;
D = FILTER C BY (B::a1 is null);
E = FOREACH D GENERATE A::a1 AS a1,A::a2 AS a2;
DUMP E;

12.每种组合有多少个
1 9
2 4
1 9
2 4

A = LOAD '/lzt02/aa01.txt' AS (a1:int,b1:int);
B = GROUP A BY (a1,a2);
C = FOREACH B GENERATE group,COUNT(A);

13.一个字符串为null 与它为空不一定等价
B = FILTER A BY (a1 is not null AND (SIZE(a1)>0L));

14.统计一个字符串中包含的指定字符数()

B = STREAM A THROUGH `awk -F "luo" '{print NF-1}'` AS (column_count:int)