Apache > Hadoop > Pig

Pig Latin Reference Manual 1


Use this manual together with Pig Latin Reference Manual 2.

Also, be sure to review the information in the Pig Cookbook.

Pig Latin Statements

A Pig Latin statement is an operator that takes a relation as input and produces another relation as output. (This definition applies to all Pig Latin operators except LOAD and STORE which read data from and write data to the file system.) Pig Latin statements can span multiple lines and must end with a semi-colon ( ; ). Pig Latin statements are generally organized in the following manner:

  1. A LOAD statement reads data from the file system.

  2. A series of "transformation" statements process the data.

  3. A STORE statement writes output to the file system; or, a DUMP statement displays output to the screen.

Running Pig Latin

You can execute Pig Latin statements:

  • Using grunt shell or command line
  • In mapreduce mode or local mode
  • Either interactively or in batch

Note that Pig now uses Hadoop's local mode (rather than Pig's native local mode).

A few run examples are shown here; see Pig Setup for more examples.

Grunt Shell - interactive, mapreduce mode (because mapreduce mode is the default you do not need to specify)

$ pig 
... - Connecting to ...
grunt> A = load 'data';
grunt> B = ... ;

Grunt Shell - batch, local mode (see the exec and run commands)

$ pig -x local
grunt> exec myscript.pig;
grunt> run myscript.pig;

Command Line - batch, mapreduce mode

$ pig myscript.pig

Command Line - batch, local mode mode

$ pig -x local myscript.pig

In general, Pig processes Pig Latin statements as follows:

  1. First, Pig validates the syntax and semantics of all statements.

  2. Next, if Pig encounters a DUMP or STORE, Pig will execute the statements.

In this example Pig will validate, but not execute, the LOAD and FOREACH statements.

A = LOAD 'student' USING PigStorage() AS (name:chararray, age:int, gpa:float);

In this example, Pig will validate and then execute the LOAD, FOREACH, and DUMP statements.

A = LOAD 'student' USING PigStorage() AS (name:chararray, age:int, gpa:float);

See Multi-Query Execution for more information on how Pig Latin statements are processed.

Pig Latin Scripts

See the following:

Pig supports running scripts that are stored in HDFS, Amazon S3, or other distributed file systems (also see REGISTER for information about Jar files). The script's full location URI is required. For example, to run a Pig script on HDFS, do the following:

java -cp pig.jar org.apache.pig.Main  hdfs://nn.mydomain.com:9020/myscripts/script.pig

Using Comments in Pig Latin Scripts

If you place Pig Latin statements in a script, the script can include comments.

  1. For multi-line comments use /* …. */

  2. For single line comments use --

/* myscript.pig
My script includes three simple Pig Latin Statements.

A = LOAD 'student' USING PigStorage() AS (name:chararray, age:int, gpa:float); -- load statement
B = FOREACH A GENERATE name;  -- foreach statement
DUMP B;  --dump statement

Retrieving Pig Latin Results

Pig Latin includes operators you can use to retrieve the results of your Pig Latin statements:

  1. Use the DUMP operator to display results to a screen.

  2. Use the STORE operator to write results to a file on the file system.

Storing Intermediate Data

Pig stores the intermediate data generated between MapReduce jobs in a temporary location on HDFS. This location must already exist on HDFS prior to use. This location can be configured using the pig.temp.dir property. The property's default value is "/tmp" which is the same as the hardcoded location in Pig 0.7.0 and earlier versions.

Debugging Pig Latin

Pig Latin includes operators that can help you debug your Pig Latin statements:

  1. Use the DESCRIBE operator to review the schema of a relation.

  2. Use the EXPLAIN operator to view the logical, physical, or map reduce execution plans to compute a relation.

  3. Use the ILLUSTRATE operator to view the step-by-step execution of a series of statements.

Complex Pig scripts often generate many MapReduce jobs. To help you debug a script, Pig prints a summary of the execution that shows which relations (aliases) are mapped to each MapReduce job.

JobId Maps Reduces MaxMapTime MinMapTIme AvgMapTime MaxReduceTime MinReduceTime AvgReduceTime Alias Feature Outputs
job_201004271216_12712 1 1 3 3 3 12 12 12 B,C GROUP_BY,COMBINER
job_201004271216_12713 1 1 3 3 3 12 12 12 D SAMPLER
job_201004271216_12714 1 1 3 3 3 12 12 12 D ORDER_BY,COMBINER hdfs://wilbur20.labs.corp.sp1.yahoo.com:9020/tmp/temp743703298/tmp-2019944040,

Working with Data

Pig Latin allows you to work with data in many ways. In general, and as a starting point:

  1. Use the FILTER operator to work with tuples or rows of data. Use the FOREACH operator to work with columns of data.

  2. Use the GROUP operator to group data in a single relation. Use the COGROUP and JOIN operators to group or join data in two or more relations.

  3. Use the UNION operator to merge the contents of two or more relations. Use the SPLIT operator to partition the contents of a relation into multiple relations.

Case Sensitivity

The names (aliases) of relations and fields are case sensitive. The names of Pig Latin functions are case sensitive. The names of parameters (see Parameter Substitution) and all other Pig Latin keywords are case insensitive.

In the example below, note the following:

  1. The names (aliases) of relations A, B, and C are case sensitive.

  2. The names (aliases) of fields f1, f2, and f3 are case sensitive.

  3. Function names PigStorage and COUNT are case sensitive.

  4. Keywords LOAD, USING, AS, GROUP, BY, FOREACH, GENERATE, and DUMP are case insensitive. They can also be written as load, using, as, group, by, etc.

  5. In the FOREACH statement, the field in relation B is referred to by positional notation ($0).

grunt> A = LOAD 'data' USING PigStorage() AS (f1:int, f2:int, f3:int);
grunt> B = GROUP A BY f1;
grunt> DUMP C;

Dynamic Invokers

Often you may need to use a simple function that is already provided by standard Java libraries, but for which a UDF has not been written. Dynamic Invokers allow you to refer to Java functions without having to wrap them in custom Pig UDFs, at the cost of doing some Java reflection on every function call.

DEFINE UrlDecode InvokeForString('java.net.URLDecoder.decode', 'String String'); 
encoded_strings = LOAD 'encoded_strings.txt' as (encoded:chararray); 
decoded_strings = FOREACH encoded_strings GENERATE UrlDecode(encoded, 'UTF-8'); 

Currently, Dynamic Invokers can be used for any static function that accepts no arguments or some combination of strings, ints, longs, doubles, floats, or arrays of same, and returns a string, an int, a long, a double, or a float. Primitives only for the numbers, no capital-letter numeric classes as arguments. Depending on the return type, a specific kind of Invoker must be used: InvokeForString, InvokeForInt, InvokeForLong, InvokeForDouble, or InvokeForFloat.

The DEFINE statement is used to bind a keyword to a Java method, as above. The first argument to the InvokeFor* constructor is the full path to the desired method. The second argument is a space-delimited ordered list of the classes of the method arguments. This can be omitted or an empty string if the method takes no arguments. Valid class names are string, long, float, double, and int. Invokers can also work with array arguments, represented in Pig as DataBags of single-tuple elements. Simply refer to string[], for example. Class names are not case-sensitive.

The ability to use invokers on methods that take array arguments makes methods like those in org.apache.commons.math.stat.StatUtils available (for processing the results of grouping your datasets, for example). This is helpful, but a word of caution: the resulting UDF will not be optimized for Hadoop, and the very significant benefits one gains from implementing the Algebraic and Accumulative interfaces are lost here. Be careful if you use invokers this way.

Memory Management

Pig allocates a fix amount of memory to store bags and spills to disk as soon as the memory limit is reached. This is very similar to how Hadoop decides when to spill data accumulated by the combiner.

The amount of memory allocated to bags is determined by pig.cachedbag.memusage; the default is set to 10% of available memory. Note that this memory is shared across all large bags used by the application.

Multi-Query Execution

With multi-query execution Pig processes an entire script or a batch of statements at once.

Turning it On or Off

Multi-query execution is turned on by default. To turn it off and revert to Pig's "execute-on-dump/store" behavior, use the "-M" or "-no_multiquery" options.

To run script "myscript.pig" without the optimization, execute Pig as follows:

$ pig -M myscript.pig
$ pig -no_multiquery myscript.pig

How it Works

Multi-query execution introduces some changes:

  1. For batch mode execution, the entire script is first parsed to determine if intermediate tasks can be combined to reduce the overall amount of work that needs to be done; execution starts only after the parsing is completed (see the EXPLAIN operator and the exec and run commands).

  2. Two run scenarios are optimized, as explained below: explicit and implicit splits, and storing intermediate results.

Explicit and Implicit Splits

There might be cases in which you want different processing on separate parts of the same data stream.

Example 1:

A = LOAD ...
SPLIT A' INTO B IF ..., C IF ...
STORE B' ...
STORE C' ...

Example 2:

A = LOAD ...
B = FILTER A' ...
C = FILTER A' ...
STORE B' ...
STORE C' ...

In prior Pig releases, Example 1 will dump A' to disk and then start jobs for B' and C'. Example 2 will execute all the dependencies of B' and store it and then execute all the dependencies of C' and store it. Both are equivalent, but the performance will be different.

Here's what the multi-query execution does to increase the performance:

  1. For Example 2, adds an implicit split to transform the query to Example 1. This eliminates the processing of A' multiple times.

  2. Makes the split non-blocking and allows processing to continue. This helps reduce the amount of data that has to be stored right at the split.

  3. Allows multiple outputs from a job. This way some results can be stored as a side-effect of the main job. This is also necessary to make the previous item work.

  4. Allows multiple split branches to be carried on to the combiner/reducer. This reduces the amount of IO again in the case where multiple branches in the split can benefit from a combiner run.

Storing Intermediate Results

Sometimes it is necessary to store intermediate results.

A = LOAD ...

If the script doesn't re-load A' for the processing of A the steps above A' will be duplicated. This is a special case of Example 2 above, so the same steps are recommended. With multi-query execution, the script will process A and dump A' as a side-effect.

Store vs. Dump

With multi-query exection, you want to use STORE to save (persist) your results. You do not want to use DUMP as it will disable multi-query execution and is likely to slow down execution. (If you have included DUMP statements in your scripts for debugging purposes, you should remove them.)

DUMP Example: In this script, because the DUMP command is interactive, the multi-query execution will be disabled and two separate jobs will be created to execute this script. The first job will execute A > B > DUMP while the second job will execute A > B > C > STORE.

A = LOAD 'input' AS (x, y, z);
B = FILTER A BY x > 5;
STORE C INTO 'output';

STORE Example: In this script, multi-query optimization will kick in allowing the entire script to be executed as a single job. Two outputs are produced: output1 and output2.

A = LOAD 'input' AS (x, y, z);
B = FILTER A BY x > 5;
STORE B INTO 'output1';
STORE C INTO 'output2';	

Error Handling

With multi-query execution Pig processes an entire script or a batch of statements at once. By default Pig tries to run all the jobs that result from that, regardless of whether some jobs fail during execution. To check which jobs have succeeded or failed use one of these options.

First, Pig logs all successful and failed store commands. Store commands are identified by output path. At the end of execution a summary line indicates success, partial failure or failure of all store commands.

Second, Pig returns different code upon completion for these scenarios:

  1. Return code 0: All jobs succeeded

  2. Return code 1: Used for retrievable errors

  3. Return code 2: All jobs have failed

  4. Return code 3: Some jobs have failed

In some cases it might be desirable to fail the entire script upon detecting the first failed job. This can be achieved with the "-F" or "-stop_on_failure" command line flag. If used, Pig will stop execution when the first failed job is detected and discontinue further processing. This also means that file commands that come after a failed store in the script will not be executed (this can be used to create "done" files).

This is how the flag is used:

$ pig -F myscript.pig
$ pig -stop_on_failure myscript.pig

Backward Compatibility

Most existing Pig scripts will produce the same result with or without the multi-query execution. There are cases though where this is not true. Path names and schemes are discussed here.

Any script is parsed in it's entirety before it is sent to execution. Since the current directory can change throughout the script any path used in LOAD or STORE statement is translated to a fully qualified and absolute path.

In map-reduce mode, the following script will load from "hdfs://<host>:<port>/data1" and store into "hdfs://<host>:<port>/tmp/out1".

cd /;
A = LOAD 'data1';
cd tmp;
STORE A INTO 'out1';

These expanded paths will be passed to any LoadFunc or Slicer implementation. In some cases this can cause problems, especially when a LoadFunc/Slicer is not used to read from a dfs file or path (for example, loading from an SQL database).

Solutions are to either:

  1. Specify "-M" or "-no_multiquery" to revert to the old names

  2. Specify a custom scheme for the LoadFunc/Slicer

Arguments used in a LOAD statement that have a scheme other than "hdfs" or "file" will not be expanded and passed to the LoadFunc/Slicer unchanged.

In the SQL case, the SQLLoader function is invoked with 'sql://mytable'.

A = LOAD 'sql://mytable' USING SQLLoader();

Implicit Dependencies

If a script has dependencies on the execution order outside of what Pig knows about, execution may fail.


In this script, MYUDF might try to read from out1, a file that A was just stored into. However, Pig does not know that MYUDF depends on the out1 file and might submit the jobs producing the out2 and out1 files at the same time.

STORE A INTO 'out1';
B = LOAD 'data2';
STORE C INTO 'out2';

To make the script work (to ensure that the right execution order is enforced) add the exec statement. The exec statement will trigger the execution of the statements that produce the out1 file.

STORE A INTO 'out1';
B = LOAD 'data2';
STORE C INTO 'out2';


In this script, the STORE/LOAD operators have different file paths; however, the LOAD operator depends on the STORE operator.

A = LOAD '/user/xxx/firstinput' USING PigStorage();
B = group ....
C = .... agrregation function
STORE C INTO '/user/vxj/firstinputtempresult/days1';
Atab = LOAD '/user/xxx/secondinput' USING  PigStorage();
Btab = group ....
Ctab = .... agrregation function
STORE Ctab INTO '/user/vxj/secondinputtempresult/days1';
E = LOAD '/user/vxj/firstinputtempresult/' USING  PigStorage();
F = group ....
G = .... aggregation function
STORE G INTO '/user/vxj/finalresult1';

Etab =LOAD '/user/vxj/secondinputtempresult/' USING  PigStorage();
Ftab = group ....
Gtab = .... aggregation function
STORE Gtab INTO '/user/vxj/finalresult2';

To make the script works, add the exec statement.

A = LOAD '/user/xxx/firstinput' USING PigStorage();
B = group ....
C = .... agrregation function
STORE C INTO '/user/vxj/firstinputtempresult/days1';
Atab = LOAD '/user/xxx/secondinput' USING  PigStorage();
Btab = group ....
Ctab = .... agrregation function
STORE Ctab INTO '/user/vxj/secondinputtempresult/days1';


E = LOAD '/user/vxj/firstinputtempresult/' USING  PigStorage();
F = group ....
G = .... aggregation function
STORE G INTO '/user/vxj/finalresult1';
Etab =LOAD '/user/vxj/secondinputtempresult/' USING  PigStorage();
Ftab = group ....
Gtab = .... aggregation function
STORE Gtab INTO '/user/vxj/finalresult2';

Optimization Rules

Pig supports various optimization rules. By default optimization, and all optimization rules, are turned on. To turn off optimiztion, use:

pig -optimizer_off [opt_rule | all ]

Note that some rules are mandatory and cannot be turned off.


Status: Mandatory

SPLIT is the only operator that models multiple outputs in Pig. To ease the process of building logical plans, all operators are allowed to have multiple outputs. As part of the optimization, all non-split operators that have multiple outputs are altered to have a SPLIT operator as the output and the outputs of the operator are then made outputs of the SPLIT operator. An example will illustrate the point. Here, a split will be inserted after the LOAD and the split outputs will be connected to the FILTER (b) and the COGROUP (c).

A = LOAD 'input';
B = FILTER A BY $1 == 1;
C = COGROUP A BY $0, B BY $0;


This rule contains several types of simplifications.

1) Constant pre-calculation 

B = FILTER A BY a0 > 5+7; 
is simplified to 
B = FILTER A BY a0 > 12; 

2) Elimination of negations 

B = FILTER A BY NOT (NOT(a0 > 5) OR a > 10); 
is simplified to 
B = FILTER A BY a0 > 5 AND a <= 10; 

3) Elimination of logical implied expression in AND 

B = FILTER A BY (a0 > 5 AND a0 > 7); 
is simplified to 
B = FILTER A BY a0 > 7; 

4) Elimination of logical implied expression in OR 

B = FILTER A BY ((a0 > 5) OR (a0 > 6 AND a1 > 15); 
is simplified to 
B = FILTER C BY a0 > 5; 

5) Equivalence elimination 

B = FILTER A BY (a0 v 5 AND a0 > 5); 
is simplified to 
B = FILTER A BY a0 > 5; 

6) Elimination of complementary expressions in OR 

B = FILTER A BY (a0 > 5 OR a0 <= 5); 
is simplified to non-filtering 

7) Elimination of naive TRUE expression 

B = FILTER A BY 1==1; 
is simplified to non-filtering 


The objective of this rule is to merge together two feach statements, if these preconditions are met:

  • The foreach statements are consecutive.
  • The first foreach statement does not contain flatten.
  • The second foreach is not nested.
-- Original code: 

A = LOAD 'file.txt' AS (a, b, c); 
B = FOREACH A GENERATE a+b AS u, c-b AS v; 

-- Optimized code: 

A = LOAD 'file.txt' AS (a, b, c); 
C = FOREACH A GENERATE a+b+5, c-b; 


The objective of this rule is to push the LIMIT operator up the data flow graph (or down the tree for database folks). In addition, for top-k (ORDER BY followed by a LIMIT) the LIMIT is pushed into the ORDER BY.

A = LOAD 'input';
B = ORDER A BY $0;
C = LIMIT B 10;


The objective of this rule is to push the FILTER operators up the data flow graph. As a result, the number of records that flow through the pipeline is reduced.

A = LOAD 'input';
B = GROUP A BY $0;
C = FILTER B BY $0 < 10;


The objective of this rule is to reduce the number of records that flow through the pipeline by moving FOREACH operators with a FLATTEN down the data flow graph. In the example shown below, it would be more efficient to move the foreach after the join to reduce the cost of the join operation.

A = LOAD 'input' AS (a, b, c);
B = LOAD 'input2' AS (x, y, z);
D = JOIN C BY $1, B BY $1;


Optimize when LOAD precedes STREAM and the loader class is the same as the serializer for the stream. Similarly, optimize when STREAM is followed by STORE and the deserializer class is same as the storage class. For both of these cases the optimization is to replace the loader/serializer with BinaryStorage which just moves bytes around and to replace the storer/deserializer with BinaryStorage.


Status: Mandatory

If you specify a schema with the LOAD statement, the optimizer will perform a pre-fix projection of the columns and cast the columns to the appropriate types. An example will illustrate the point. The LOAD statement (a) has a schema associated with it. The optimizer will insert a FOREACH operator that will project columns 0, 1 and 2 and also cast them to chararray, int and float respectively.

A = LOAD 'input' AS (name: chararray, age: int, gpa: float);
B = FILER A BY $1 == 1;
C = GROUP A By $0;

Pig Properties

The Pig "-propertyfile " option enables you to pass a set of Pig or Hadoop properties to a Pig job. If the value is present in both the property file passed from the command line as well as in default property file bundled into pig.jar, the properties passed from command line take precedence. This property, as well as all other properties defined in Pig, are available to your UDFs via UDFContext.getClientSystemProps()API call (see the Pig UDF Manual.)

You can retrieve a list of all properties using the help properties command.

You can set properties using the set command.

Pig Statistics

Pig Statistics is a framework for collecting and storing script-level statistics for Pig Latin. Characteristics of Pig Latin scripts and the resulting MapReduce jobs are collected while the script is executed. These statistics are then available for Pig users and tools using Pig (such as Oozie) to retrieve after the job is done.

The new Pig statistics and the existing Hadoop statistics can also be accessed via the Hadoop job history file (and job xml file). Piggybank has a HadoopJobHistoryLoader which acts as an example of using Pig itself to query these statistics (the loader can be used as a reference implementation but is NOT supported for production use).

Java API

Several new public classes make it easier for external tools such as Oozie to integrate with Pig statistics.

The Pig statistics are available here: http://pig.apache.org/docs/r0.8.1/api/

The stats classes are in the package: org.apache.pig.tools.pigstats

  • PigStats
  • JobStats
  • OutputStats
  • InputStats

The PigRunner class mimics the behavior of the Main class but gives users a statistics object back. Optionally, you can call the API with an implementation of progress listener which will be invoked by Pig runtime during the execution.

package org.apache.pig;

public abstract class PigRunner {
    public static PigStats run(String[] args, PigProgressNotificationListener listener)

public interface PigProgressNotificationListener extends java.util.EventListener {
    // just before the launch of MR jobs for the script
    public void LaunchStartedNotification(int numJobsToLaunch);
    // number of jobs submitted in a batch
    public void jobsSubmittedNotification(int numJobsSubmitted);
    // a job is started
    public void jobStartedNotification(String assignedJobId);
    // a job is completed successfully
    public void jobFinishedNotification(JobStats jobStats);
    // a job is failed
    public void jobFailedNotification(JobStats jobStats);
    // a user output is completed successfully
    public void outputCompletedNotification(OutputStats outputStats);
    // updates the progress as percentage
    public void progressUpdatedNotification(int progress);
    // the script execution is done
    public void launchCompletedNotification(int numJobsSucceeded);


The following entries are included in job conf:

Pig Statistic



The UUID for the script. All jobs spawned by the script have the same script ID.


The base64 encoded script text.


The command line used to invoke the script.


The Hadoop version installed.


The Pig version used.


A comma-separated list of input directories for the job.


A comma-separated list of output directories in the map phase of the job.


A comma-separated list of output directories in the reduce phase of the job.


A comma-separated list of parent job ids.


A list of Pig features used in the script.


A list of Pig features used in the job.


The alias associated with the job.

Hadoop Job History Loader

The HadoopJobHistoryLoader in Piggybank loads Hadoop job history files and job xml files from file system. For each MapReduce job, the loader produces a tuple with schema (j:map[], m:map[], r:map[]). The first map in the schema contains job-related entries. Here are some of important key names in the map:


















Examples that use the loader to query Pig statistics are shown below.


Find scripts that generate more then three MapReduce jobs:

a = load '/mapred/history/done' using HadoopJobHistoryLoader() as (j:map[], m:map[], r:map[]);
b = group a by (j#'PIG_SCRIPT_ID', j#'USER', j#'JOBNAME');
c = foreach b generate group.$1, group.$2, COUNT(a);
d = filter c by $2 > 3;
dump d;

Find the running time of each script (in seconds):

a = load '/mapred/history/done' using HadoopJobHistoryLoader() as (j:map[], m:map[], r:map[]);
b = foreach a generate j#'PIG_SCRIPT_ID' as id, j#'USER' as user, j#'JOBNAME' as script_name, 
         (Long) j#'SUBMIT_TIME' as start, (Long) j#'FINISH_TIME' as end;
c = group b by (id, user, script_name)
d = foreach c generate group.user, group.script_name, (MAX(b.end) - MIN(b.start)/1000;
dump d;

Find the number of scripts run by user and queue on a cluster:

a = load '/mapred/history/done' using HadoopJobHistoryLoader() as (j:map[], m:map[], r:map[]);
b = foreach a generate j#'PIG_SCRIPT_ID' as id, j#'USER' as user, j#'QUEUE_NAME' as queue;
c = group b by (id, user, queue) parallel 10;
d = foreach c generate group.user, group.queue, COUNT(b);
dump d;

Find scripts that have failed jobs:

a = load '/mapred/history/done' using HadoopJobHistoryLoader() as (j:map[], m:map[], r:map[]);
b = foreach a generate (Chararray) j#'STATUS' as status, j#'PIG_SCRIPT_ID' as id, j#'USER' as user, j#'JOBNAME' as script_name, j#'JOBID' as job;
c = filter b by status != 'SUCCESS';
dump c;

Find scripts that use only the default parallelism:

a = load '/mapred/history/done' using HadoopJobHistoryLoader() as (j:map[], m:map[], r:map[]);
b = foreach a generate j#'PIG_SCRIPT_ID' as id, j#'USER' as user, j#'JOBNAME' as script_name, (Long) r#'NUMBER_REDUCES' as reduces;
c = group b by (id, user, script_name) parallel 10;
d = foreach c generate group.user, group.script_name, MAX(b.reduces) as max_reduces;
e = filter d by max_reduces == 1;
dump e;

Specialized Joins

In certain cases, the performance of inner joins and outer joins can be optimized using replicated, skewed, or merge joins.

Replicated Joins

Fragment replicate join is a special type of join that works well if one or more relations are small enough to fit into main memory. In such cases, Pig can perform a very efficient join because all of the hadoop work is done on the map side. In this type of join the large relation is followed by one or more small relations. The small relations must be small enough to fit into main memory; if they don't, the process fails and an error is generated.


Perform a replicated join with the USING clause (see inner joins and outer joins). In this example, a large relation is joined with two smaller relations. Note that the large relation comes first followed by the smaller relations; and, all small relations together must fit into main memory, otherwise an error is generated.

big = LOAD 'big_data' AS (b1,b2,b3);

tiny = LOAD 'tiny_data' AS (t1,t2,t3);

mini = LOAD 'mini_data' AS (m1,m2,m3);

C = JOIN big BY b1, tiny BY t1, mini BY m1 USING 'replicated';


Fragment replicate joins are experimental; we don't have a strong sense of how small the small relation must be to fit into memory. In our tests with a simple query that involves just a JOIN, a relation of up to 100 M can be used if the process overall gets 1 GB of memory. Please share your observations and experience with us.

Skewed Joins

Parallel joins are vulnerable to the presence of skew in the underlying data. If the underlying data is sufficiently skewed, load imbalances will swamp any of the parallelism gains. In order to counteract this problem, skewed join computes a histogram of the key space and uses this data to allocate reducers for a given key. Skewed join does not place a restriction on the size of the input keys. It accomplishes this by splitting the left input on the join predicate and streaming the right input. The left input is sampled to create the histogram.

Skewed join can be used when the underlying data is sufficiently skewed and you need a finer control over the allocation of reducers to counteract the skew. It should also be used when the data associated with a given key is too large to fit in memory.


Perform a skewed join with the USING clause (see inner joins and outer joins).

big = LOAD 'big_data' AS (b1,b2,b3);
massive = LOAD 'massive_data' AS (m1,m2,m3);
C = JOIN big BY b1, massive BY m1 USING 'skewed';


Skewed join will only work under these conditions:

  • Skewed join works with two-table inner join. Currently we do not support more than two tables for skewed join. Specifying three-way (or more) joins will fail validation. For such joins, we rely on you to break them up into two-way joins.
  • The pig.skewedjoin.reduce.memusage Java parameter specifies the fraction of heap available for the reducer to perform the join. A low fraction forces pig to use more reducers but increases copying cost. We have seen good performance when we set this value in the range 0.1 - 0.4. However, note that this is hardly an accurate range. Its value depends on the amount of heap available for the operation, the number of columns in the input and the skew. An appropriate value is best obtained by conducting experiments to achieve a good performance. The default value is 0.5.
  • Skewed join does not address (balance) uneven data distribution across reducers. However, in most cases, skewed join ensures that the join will finish (however slowly) rather than fail.

Merge Joins

Often user data is stored such that both inputs are already sorted on the join key. In this case, it is possible to join the data in the map phase of a MapReduce job. This provides a significant performance improvement compared to passing all of the data through unneeded sort and shuffle phases.

Pig has implemented a merge join algorithm, or sort-merge join, although in this case the sort is already assumed to have been done (see the Conditions, below). Pig implements the merge join algorithm by selecting the left input of the join to be the input file for the map phase, and the right input of the join to be the side file. It then samples records from the right input to build an index that contains, for each sampled record, the key(s) the filename and the offset into the file the record begins at. This sampling is done in the first MapReduce job. A second MapReduce job is then initiated, with the left input as its input. Each map uses the index to seek to the appropriate record in the right input and begin doing the join.


Perform a merge join with the USING clause (see inner joins and outer joins).

C = JOIN A BY a1, B BY b1, C BY c1 USING 'merge';


Condition A

Inner merge join (between two tables) will only work under these conditions:

  • Between the load of the sorted input and the merge join statement there can only be filter statements and foreach statement where the foreach statement should meet the following conditions:
    • There should be no UDFs in the foreach statement.
    • The foreach statement should not change the position of the join keys.
    • There should be no transformation on the join keys which will change the sort order.
  • Data must be sorted on join keys in ascending (ASC) order on both sides.
  • Right-side loader must implement either the {OrderedLoadFunc} interface or {IndexableLoadFunc} interface.
  • Type information must be provided for the join key in the schema.

The Zebra and PigStorage loaders satisfy all of these conditions.

Condition B

Outer merge join (between two tables) and inner merge join (between three or more tables) will only work under these conditions:

  • No other operations can be done between the load and join statements.
  • Data must be sorted on join keys in ascending (ASC) order on both sides.
  • Left-most loader must implement {CollectableLoader} interface as well as {OrderedLoadFunc}.
  • All other loaders must implement {IndexableLoadFunc}.
  • Type information must be provided for the join key in the schema.

The Zebra loader satisfies all of these conditions.

An example of a left outer merge join using the Zebra loader:

A = load 'data1' using org.apache.hadoop.zebra.pig.TableLoader('id:int', 'sorted'); 
B = load 'data2' using org.apache.hadoop.zebra.pig.TableLoader('id:int', 'sorted'); 
C = join A by id left, B by id using 'merge'; 

Both Conditions

For optimal performance, each part file of the left (sorted) input of the join should have a size of at least 1 hdfs block size (for example if the hdfs block size is 128 MB, each part file should be less than 128 MB). If the total input size (including all part files) is greater than blocksize, then the part files should be uniform in size (without large skews in sizes). The main idea is to eliminate skew in the amount of input the final map job performing the merge-join will process.

Zebra Integration

For information about how to integrate Zebra with your Pig scripts, see Zebra and Pig.