We are seeing a number of Pig 0.8.1 unit test failures with Apache Hadoop 0.22
due to differences in the ordering of the expected results.
For example, in TestCombiner.testJiraPig746, the expected result is:
"(pig1,75,{(pig1,18,2.1),(pig1,18,2.1),(pig1,19,2.1),(pig1,20,3.1)})",
But the actual result is:
"(pig1,75,{(pig1,20,3.1),(pig1,19,2.1),(pig1,18,2.1),(pig1,18,2.1)})",
Other similar test failures include:
TestCommit.testCheckin2
TestEvalPipeline2.testOrderByLimitJoin
TestMultiQuery.testMultiQueryJiraPig1113
Are these valid test issues or is there something obvious I am missing?
Thanks,
Po