Pig -- a language for Hadoop
Post date: May 7, 2014 4:11:04 AM
Some note taken at Pig presentation by Jared May 2014
Pig – language for Hadoop
- step by step (comperative) language
- compile into Hadoop job
Hive vs Pig
Both compile into Hadoop
Hive – sql-like lang
Pig – comparative lange, do X, Y, Z, can do more complex things
SQL – not good for complex things – too many nested
Some prefer Pig over Hive with this reason
Pig:
- Tuple – row that has schema
- Bag
Load – similar to pointer, not load
Pig works for every data in MDS
to run a Pig script
/Apollo/env/RFUGE/perl-script/run-pig-script
Pig supports Python, Ruby
Hadoop’s lower bound = 1 min