[PIG-4930] Skewed Join Breaks On Empty Sampled Input When Key is From Map

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.9.2, 0.16.0
• Fix Version/s: 0.17.0, 0.16.1
• Hadoop Flags: Reviewed
• Assignee: Nándor Kollár
• Reporter: William Butler
• Created: 16/Jun/16 18:38
• Updated: 21/Jun/17 09:15
• Resolved: 07/Jan/17 20:20

Description

When using a skewed join, if the left relation gets its key from a map and said relation is empty, then the skewed join fails during the sampling phase with:

org.apache.pig.backend.executionengine.ExecException: ERROR 0: Exception while executing (Name: Local Rearrange[tuple]

{tuple}

(false) - scope-27 Operator Key: scope-27): org.apache.pig.backend.executionengine.ExecException: ERROR 0: Exception while executing [POMapLookUp (Name: POMapLookUp[bytearray] - scope-14 Operator Key: scope-14) children: null at [null[3,17]]]: java.lang.ClassCastException: java.lang.String cannot be cast to java.util.Map
at org.apache.pig.backend.hadoop.executionengine.physicalLayer.PhysicalOperator.processInput(PhysicalOperator.java:314)
at org.apache.pig.backend.hadoop.executionengine.physicalLayer.relationalOperators.POLocalRearrange.getNextTuple(POLocalRearrange.java:287)
at org.apache.pig.backend.hadoop.executionengine.mapReduceLayer.PigGenericMapBase.runPipeline(PigGenericMapBase.java:280)
at org.apache.pig.backend.hadoop.executionengine.mapReduceLayer.PigGenericMapBase.map(PigGenericMapBase.java:275)
at org.apache.pig.backend.hadoop.executionengine.mapReduceLayer.PigGenericMapBase.map(PigGenericMapBase.java:65)
at org.apache.hadoop.mapreduce.Mapper.run(Mapper.java:144)
at org.apache.hadoop.mapred.MapTask.runNewMapper(MapTask.java:764)
at org.apache.hadoop.mapred.MapTask.run(MapTask.java:370)
at org.apache.hadoop.mapred.Child$4.run(Child.java:255)
at java.security.AccessController.doPrivileged(Native Method)
at javax.security.auth.Subject.doAs(Subject.java:415)
at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1121)
at org.apache.hadoop.mapred.Child.main(Child.java:249)

I think the problem is more fundamental to Pig's skewed join implementation than maps, but it is easily demonstrable with them. I have written an additional test in TestSkewedJoin that demonstrates the problem. The join works correctly if we remove "using 'skewed'"