[IMPALA-2080] Automatically gather row count stats during Create Table As Select

• Type: Improvement
• Status: Resolved
• Priority: Minor
• Resolution: Won't Fix
• Affects Version/s: Impala 2.2
• Component/s: Backend
• Labels: ramp-up
• Target Version: Product Backlog
• Assignee: Unassigned
• Reporter: Alan Choi
• Created: 18/Jun/15 18:57
• Updated: 14/Mar/17 00:53
• Resolved: 14/Mar/17 00:53

Description

Problem Statement
During ETL process or ad-hoc analysis, a temporary table is created using Create Table As Select (CTAS) and then read immediately by the subsequent query. However, the newly create temp table has no stats. Subsequent queries that read from the temp table can have an incorrect join order and join strategy.

Potential Fix
Given that row count is readily avaialble during CTAS and it's the most important fact in determining join order and join strategy, Impala should update the row count stats of the table created during CTAS.