danharvey / pigjsonutils Goto Github PK

= Overview
Pig StoreFunc to save Pig tuples into rows of JSON objects.

The schema is derevied from the Pig schema  and supports all numeric, string, boolean types along with Bags and embeded Tuples.
Not all Pig types are supported in JSON so not all tuples can be serialised with this, notebly the raw byte types.
The Map type could also be supported but is not currently coded.

== Usage

To use you need to register the jar then you can use the loader and normal :-

REGISTER lib/json-simple-1.1.jar;
A = ...
STORE A INTO 'test.json' USING uk.co.danharvey.pig.storage.JsonStorage();

== ToDo

The schema is currently shared between the output tasks using the same hdfs folder as the output using a file called _schema.json so this probably needs remove to make it behave the same as other map reduce jobs and output formats.