[ https://issues.apache.org/jira/browse/SPARK-12313?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Harsh Gupta updated SPARK-12313: -------------------------------- Comment: was deleted (was: [~liancheng] Should this issue occur after your PR https://github.com/apache/spark/pull/7492 ?) > getPartitionsByFilter doesnt handle predicates on all / multiple Partition > Columns > ---------------------------------------------------------------------------------- > > Key: SPARK-12313 > URL: https://issues.apache.org/jira/browse/SPARK-12313 > Project: Spark > Issue Type: Bug > Components: SQL > Affects Versions: 1.5.1 > Reporter: Gobinathan SP > Assignee: Harsh Gupta > Priority: Critical > > When enabled spark.sql.hive.metastorePartitionPruning, the > getPartitionsByFilter is used > For a table partitioned by p1 and p2, when triggered hc.sql("select col.... > from tabl1 where p1='p1V' and p2= 'p2V' ") > The HiveShim identifies the Predicates and ConvertFilters returns p1='p1V' > and col2= 'p2V' . The same is passed to the getPartitionsByFilter method as > filter string. > On these cases the partitions are not returned from Hive's > getPartitionsByFilter method. As a result, for the sql, the number of > returned rows is always zero. > However, filter on a single column always works. Probalbly it doesn't come > through this route > I'm using Oracle for Metstore V0.13.1 -- This message was sent by Atlassian JIRA (v6.3.4#6332) --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org