Pig Small Case 1

Premise: If there is a batch of input data format as follows:

User id| Article category | views

requirements: use pig to find out the most visited article category for each user?

1. First: read in the data

– Specify the delimiter as "|"
– Specify the type of each field

Enter pig, u_ct=load '/output ' using Pigstorage (' | ') as (user:chararray,category:chararray,times:int);

2. Sub-user statistics:
–group: Aggregation of the same user's memory login
–sorted: Sort from large to small according to Times
–top: Keep The first line of sorted

U_stat=foreach (group u_ct by user) {sorted= order u_ct by times desc; top= limit sorted 1; Generate flatten (top), SUM (u_ct. Times)};

3. Output: Store u_stat into '/stat ';

4 Show: Cat/stat

Pig Small Case 1