Uploaded image for project: 'Pig'
  1. Pig
  2. PIG-2119

DuplicateForEachColumnRewrite makes assumptions about the position of LOGGenerate in the plan

    Details

    • Type: Bug
    • Status: Closed
    • Priority: Major
    • Resolution: Fixed
    • Affects Version/s: 0.9.0, 0.9.1
    • Fix Version/s: 0.9.2, 0.10.0, 0.11
    • Component/s: None
    • Labels:
      None
    • Hadoop Flags:
      Reviewed

      Description

      The input:

      grunt> cat b.txt
      a       11
      b       3
      c       10
      a       12
      b       10
      c       15
      

      The script:

      a = load 'b.txt' AS (id:chararray, num:int);
      b = group a by id;
      c = foreach b { 
        d = order a by num DESC;
        n = COUNT(a);
        e = limit d 1;
        generate n;
      }
      

      The exception:

      Caused by: java.lang.ClassCastException: org.apache.pig.newplan.logical.relational.LOLimit cannot be cast to org.apache.pig.newplan.logical.relational.LOGenerate
              at org.apache.pig.newplan.logical.rules.DuplicateForEachColumnRewrite$DuplicateForEachColumnRewriteTransformer.check(DuplicateForEachColumnRewrite.java:87)
              at org.apache.pig.newplan.optimizer.PlanOptimizer.optimize(PlanOptimizer.java:108)
      
      

      I know the script is a bit pointless, but I was just testing and modifying the script bit by bit.
      If I remove the limit in any case I get the same exception but with LOSort.

      The problem, I think, is that the rule assumes there is only 1 sink in the nested block and that this sink is a LOGenerate.

        Activity

        Hide
        daijy Daniel Dai added a comment -

        LOGenerate is the only sink in nested plan. This is one of the basic assumption we made throughout logical plan. To solve the issue above, we need to prune the dangling branch before proceed.

        Show
        daijy Daniel Dai added a comment - LOGenerate is the only sink in nested plan. This is one of the basic assumption we made throughout logical plan. To solve the issue above, we need to prune the dangling branch before proceed.
        Hide
        vivekp Vivek Padmanabhan added a comment -

        Faced this issue with the below script;

        A = load '3char_1long_tab' as (f1:chararray, f2:chararray, f3:chararray,ct:long);
        B = GROUP A  BY f1;
        C =    FOREACH B {
                zip_ordered = ORDER A BY f3 ASC; 
                GENERATE
                        FLATTEN(group) AS f1,	
                        A.(f3, ct),
        		--COUNT(zip_ordered),
                        SUM(A.ct) AS total;
          };
        
        dump C;
        

        The zip_ordered is an accident and not used, but Pig 0.8 silently ignores this while Pig 0.9 throws exception.
        I believe the affect version should be 0.9

        Show
        vivekp Vivek Padmanabhan added a comment - Faced this issue with the below script; A = load '3char_1long_tab' as (f1:chararray, f2:chararray, f3:chararray,ct: long ); B = GROUP A BY f1; C = FOREACH B { zip_ordered = ORDER A BY f3 ASC; GENERATE FLATTEN(group) AS f1, A.(f3, ct), --COUNT(zip_ordered), SUM(A.ct) AS total; }; dump C; The zip_ordered is an accident and not used, but Pig 0.8 silently ignores this while Pig 0.9 throws exception. I believe the affect version should be 0.9
        Hide
        thejas Thejas M Nair added a comment -

        +1

        Show
        thejas Thejas M Nair added a comment - +1
        Hide
        daijy Daniel Dai added a comment -

        Unit tests pass. Test-patch:
        [exec] -1 overall.
        [exec]
        [exec] +1 @author. The patch does not contain any @author tags.
        [exec]
        [exec] +1 tests included. The patch appears to include 6 new or modified tests.
        [exec]
        [exec] +1 javadoc. The javadoc tool did not generate any warning messages.
        [exec]
        [exec] +1 javac. The applied patch does not increase the total number of javac compiler warnings.
        [exec]
        [exec] +1 findbugs. The patch does not introduce any new Findbugs warnings.
        [exec]
        [exec] -1 release audit. The applied patch generated 458 release audit warnings (more than the trunk's current 447 warnings).

        All new files has proper header.

        Patch committed to both trunk and 0.10 branch.

        Show
        daijy Daniel Dai added a comment - Unit tests pass. Test-patch: [exec] -1 overall. [exec] [exec] +1 @author. The patch does not contain any @author tags. [exec] [exec] +1 tests included. The patch appears to include 6 new or modified tests. [exec] [exec] +1 javadoc. The javadoc tool did not generate any warning messages. [exec] [exec] +1 javac. The applied patch does not increase the total number of javac compiler warnings. [exec] [exec] +1 findbugs. The patch does not introduce any new Findbugs warnings. [exec] [exec] -1 release audit. The applied patch generated 458 release audit warnings (more than the trunk's current 447 warnings). All new files has proper header. Patch committed to both trunk and 0.10 branch.
        Hide
        daijy Daniel Dai added a comment -

        Patch committed to 0.9 branch as per Dmitriy's request (PIG-2474)

        Show
        daijy Daniel Dai added a comment - Patch committed to 0.9 branch as per Dmitriy's request ( PIG-2474 )

          People

          • Assignee:
            daijy Daniel Dai
            Reporter:
            azaroth Gianmarco De Francisci Morales
          • Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

            Dates

            • Created:
              Updated:
              Resolved:

              Development