[ https://issues.apache.org/jira/browse/HIVE-18695?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16377497#comment-16377497 ]
Josh Elser commented on HIVE-18695: ----------------------------------- bq. From the comments above, it seems like a AccumuloStorageHandler bug where it seems to depend on PPD having been done ( I haven't done a deeper investigation, which would be necessary to confirm). If memory serves me, yes, the storage handler is expecting that PPD (which I'm reading as predicate pushdown, but am honestly not positive that's what the acronym is ;)) happens. Accumulo has a server-side computation layer in which we take some operands from the Hive query and can push them into Accumulo for efficiency's sake. Nothing more than filters is happening now (e.g. {{key == 1}} and {{key == 2}} in the above query), again, if memory serves. Is there some new kind of API which the storage handlers can look at to figure out when PPD is going to happen (or not)? That would have serious correctness issues if not.. [~erwaman], I'm confused on the phrasing: does HIVE-18802 apply both before and after HIVE-15680? I think you're saying that it was an existing problem, but it would also fix the test failures that HIVE-15680 caused? > fix TestAccumuloCliDriver.testCliDriver[accumulo_queries] > --------------------------------------------------------- > > Key: HIVE-18695 > URL: https://issues.apache.org/jira/browse/HIVE-18695 > Project: Hive > Issue Type: Bug > Reporter: Zoltan Haindrich > Priority: Major > > seems to be broken by HIVE-15680 -- This message was sent by Atlassian JIRA (v7.6.3#76005)