Pig: Introduction to Latin - 1

• Relation and Field

Pig Latin is a dataflow language. Each processing step results in a new data set, or relation.

A = load 'NYSE_dividends' (exchange, symbol, date, dividends);

//A is relation      exchange,symbol,date and dividends are all fields

 

• Case Sensitivity

Keywords in Pig Latin are not case-sensitive; for example, LOAD is equivalent to load. But relation and field names are.UDF names are also case-sensitive, thus COUNT is not the same UDF as count.

 

• Comments

A = load 'foo'; --this is a single-line comment
/*
* This is a multiline comment.
*/
B = load /* a comment in the middle */'bar';

 

• Load

PigStorage and TextLoader

divs = load '/data/examples/NYSE_dividends';      //tab-delimited file

divs = load 'NYSE_dividends' using HBaseStorage();  //load from hbase

divs = load 'NYSE_dividends' using PigStorage(',');    //comma-separated text data

divs = load 'NYSE_dividends' as (exchange, symbol, date, dividends);  //specify the schema

divs = load 'datadir'; // read all files recursively in the datadir

divs = load 'datadir/part-2012-*';  //read multiple files in datadir

 

• Store

store processed into '/data/examples/processed';

store processed into 'processed' using HBaseStorage();

store processed into 'processed' using PigStorage(',');

 

Note:when writing to a filesystem, processed will be a directory with part files rather than a single file. But how many part files will be created? That depends on the parallelism of the last job before the store. If it has re-duces, it will be determined by the parallel level set for that job.  If it is a map-only job, it will be determined by the number of maps, which is controlled by Hadoop and not Pig.

 

• Dump

dump processed;  //sent contents of processed to console