[HIVE-3602] Provide ability to roll up column stats from specific partitions

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 0.10.0
• Component/s: Statistics
• Assignee: Shreepadma Venugopalan
• Reporter: Shreepadma Venugopalan
• Created: 20/Oct/12 18:07
• Updated: 02/May/13 02:29

Description

When executing query on a partitioned table it becomes necessary to combine the statistics from different partitions to know about the combined data distribution. For eg., consider a table with 5 partitions P0 to P4 and a query that scans partitions p0, p2 and p4. It becomes necessary to combine the statistics from the individual partitions to obtain the statistical properties of the combined data from the partitions. This JIRA covers the task of a) implementing logic to roll up statistics from specific partitions and b) providing an API to retrieve the combined statistics

Issue Links

• depends upon: HIVE-1362 Optimizer statistics on columns in tables and partitions
• depends upon: HIVE-3516 Fast incremental statistics computation on columns in Hive tables