Pig Interpreter for Apache Zeppelin

Overview

Apache Pig is a platform for analyzing large data sets that consists of a high-level language for expressing data analysis programs, coupled with infrastructure for evaluating these programs. The salient property of Pig programs is that their structure is amenable to substantial parallelization, which in turns enables them to handle very large data sets.

Supported interpreter type

  • %pig.script (default)

    All the pig script can run in this type of interpreter, and display type is plain text.

  • %pig.query

    Almost the same as %pig.script. The only difference is that you don't need to add alias in the last statement. And the display type is table.

Supported runtime mode

  • Local
  • MapReduce
  • Tez_Local (Only Tez 0.7 is supported)
  • Tez (Only Tez 0.7 is supported)

How to use

How to setup Pig

  • Local Mode

    Nothing needs to be done for local mode

  • MapReduce Mode

    HADOOP_CONF_DIR needs to be specified in ZEPPELIN_HOME/conf/zeppelin-env.sh.

  • Tez Local Mode

    Nothing needs to be done for tez local mode

  • Tez Mode

    HADOOP_CONF_DIR and TEZ_CONF_DIR needs to be specified in ZEPPELIN_HOME/conf/zeppelin-env.sh.

How to configure interpreter

At the Interpreters menu, you have to create a new Pig interpreter. Pig interpreter has below properties by default. And you can set any pig properties here which will be passed to pig engine. (like tez.queue.name & mapred.job.queue.name). Besides, we use paragraph title as job name if it exists, else use the last line of pig script. So you can use that to find app running in YARN RM UI.

Property Default Description
zeppelin.pig.execType mapreduce Execution mode for pig runtime. local | mapreduce | tez_local | tez
zeppelin.pig.includeJobStats false whether display jobStats info in %pig.script
zeppelin.pig.maxResult 1000 max row number displayed in %pig.query
tez.queue.name default queue name for tez engine
mapred.job.queue.name default queue name for mapreduce engine

Example

pig
%pig

raw_data = load 'dataset/sf_crime/train.csv' using PigStorage(',') as (Dates,Category,Descript,DayOfWeek,PdDistrict,Resolution,Address,X,Y);
b = group raw_data all;
c = foreach b generate COUNT($1);
dump c;
pig.query
%pig.query

b = foreach raw_data generate Category;
c = group b by Category;
foreach c generate group as category, COUNT($1) as count;

Data is shared between %pig and %pig.query, so that you can do some common work in %pig, and do different kinds of query based on the data of %pig. Besides, we recommend you to specify alias explicitly so that the visualization can display the column name correctly. Here, we name COUNT($1) as count, if you don't do this, then we will name it using position, here we will use col_1 to represent COUNT($1) if you don't specify alias for it. There's one pig tutorial note in zeppelin for your reference.