Author: knoguchi Date: Wed Aug 12 17:07:10 2015 New Revision: 1695579 URL: http://svn.apache.org/r1695579 Log: PIG-4628: Pig 0.14 job with order by fails in mapreduce mode with Oozie (knoguchi)
Modified: pig/branches/branch-0.15/CHANGES.txt pig/branches/branch-0.15/src/org/apache/pig/backend/hadoop/executionengine/mapReduceLayer/JobControlCompiler.java Modified: pig/branches/branch-0.15/CHANGES.txt URL: http://svn.apache.org/viewvc/pig/branches/branch-0.15/CHANGES.txt?rev=1695579&r1=1695578&r2=1695579&view=diff ============================================================================== --- pig/branches/branch-0.15/CHANGES.txt (original) +++ pig/branches/branch-0.15/CHANGES.txt Wed Aug 12 17:07:10 2015 @@ -28,6 +28,8 @@ OPTIMIZATIONS BUG FIXES +PIG-4628: Pig 0.14 job with order by fails in mapreduce mode with Oozie (knoguchi) + PIG-4627: [Pig on Tez] Self join does not handle null values correctly (rohini) PIG-4649: [Pig on Tez] Union followed by HCatStorer misses some data (rohini) Modified: pig/branches/branch-0.15/src/org/apache/pig/backend/hadoop/executionengine/mapReduceLayer/JobControlCompiler.java URL: http://svn.apache.org/viewvc/pig/branches/branch-0.15/src/org/apache/pig/backend/hadoop/executionengine/mapReduceLayer/JobControlCompiler.java?rev=1695579&r1=1695578&r2=1695579&view=diff ============================================================================== --- pig/branches/branch-0.15/src/org/apache/pig/backend/hadoop/executionengine/mapReduceLayer/JobControlCompiler.java (original) +++ pig/branches/branch-0.15/src/org/apache/pig/backend/hadoop/executionengine/mapReduceLayer/JobControlCompiler.java Wed Aug 12 17:07:10 2015 @@ -697,6 +697,11 @@ public class JobControlCompiler{ conf.set("pig.inpTargets", ObjectSerializer.serialize(inpTargets)); conf.set("pig.inpSignatures", ObjectSerializer.serialize(inpSignatureLists)); conf.set("pig.inpLimits", ObjectSerializer.serialize(inpLimits)); + + // Removing job credential entry before serializing pigcontext into jobconf + // since this path would be invalid for the new job being created + pigContext.getProperties().remove("mapreduce.job.credentials.binary"); + conf.set("pig.pigContext", ObjectSerializer.serialize(pigContext)); conf.set("udf.import.list", ObjectSerializer.serialize(PigContext.getPackageImportList())); // this is for unit tests since some don't create PigServer