[HIVE-3526] Column Statistics - Add support for equi-height histograms on numeric columns

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 0.10.0
• Component/s: Statistics
• Assignee: Shreepadma Venugopalan
• Reporter: Shreepadma Venugopalan
• Created: 03/Oct/12 22:09
• Updated: 06/Nov/12 21:21

Description

This JIRA covers the task of adding support for equi-height histograms on numeric columns in Hive tables and partitions. This task involves a) implementing a UDAF to compute equi-height histograms on numeric columns, b) persisting the histogram to the metastore along with other column statistics , c) enhancing the thrift api to retrieve the histogram along with the other statistics and d) extending the grammar of ANALYZE to allow the user to request histograms and specify the number of bins.

Issue Links

• is related to: HIVE-1362 Optimizer statistics on columns in tables and partitions

Sub-Tasks

• Implement equi-depth histograms as a UDAF: Open
• Enhance the existing thrift APIs to persist the histogram: Open
• Enhance the existing thrift APIs to retrieve the histogram corresponding to a column: Open
• Extend the analyze table syntax to allow the user to specify the histogram bins: Open
• Extend the analyze table syntax to allow the user to request computing histogram: Open