Details
-
Sub-task
-
Status: Resolved
-
Major
-
Resolution: Duplicate
-
None
-
None
Description
Union_3 e2e test is failing on spark branch. If I execute on spark branch in MR mode (test-e2e instead of test-e2e-spark), the test is passing.
It fails while comparing the output of the second foreach (the output of the first foreach is fine):
RESULTS DIFFER: vimdiff /var/lib/hadoop-hdfs/pig-spark/pig/test/e2e/pig/testdist/./out/pigtest/hdfs/hdfs-1487068760-nightly.conf/Union_3.out/2/out_sorted ./out/pigtest/hdfs/hdfs-1487068760-nightly.conf/Union_3_benchmark.out/2/out_sorted
The first 3 lines:
diff <(head -n 3 ./Union_3_benchmark.out/2/out_sorted) <(head -n 3 ./Union_3.out/2/out_sorted) 1,3c1,3 < {(alice allen,19,1.93)} < {(alice brown,18,2.35)} < {(alice carson,19,1.88)} --- > {(alice allen,77,3.51)} > {(alice brown,77,0.74)} > {(alice carson,74,1.73)}
Old Pig is path is ../../../.., so the benchmarks were created with the same Pig in MR mode.
Attachments
Issue Links
- is duplicated by
-
PIG-5171 SecondarySort_7 is failing with spark exec type
- Closed