[SPARK-6784] Make sure values of partitioning columns are correctly converted based on their data types

• Type: Bug
• Status: Resolved
• Priority: Blocker
• Resolution: Duplicate
• Component/s: SQL
• Assignee: Adrian Wang
• Reporter: Davies Liu
• Created: 08/Apr/15 21:54
• Updated: 25/May/15 09:31
• Resolved: 25/May/15 09:31

Description

We used to have the problems that values of partitioning columns are not correctly cast to the desired Spark SQL values based on their data types. Let's make sure we correctly do that for both Hive's partitions and HadoopFSRelation's partitions.

Issue Links

• contains: SPARK-5456 Decimal Type comparison issue
• duplicates: SPARK-7790 when use dynamic partitions, the partition string can be wrong without looking at the type
• links to: [Github] Pull Request #6027 (scwf)