Performance improvement for queries having join between TABLE_CONSTRAINTS and KEY_COLUMN_USAGE views

[sumitj824]

Description

This commit addresses poor performance issues in queries that join the TABLE_CONSTRAINTS and KEY_COLUMN_USAGE views. The problem was caused by inaccurate row estimates resulting from a CASE expression in the TABLE_CONSTRAINTS view, leading to suboptimal query plans.

Signed-off-by: Sumit Jaiswal sumiji@amazon.com

Issues Resolved

Task: BABEL-5427

Test Scenarios Covered

Test are already present for modified views.

Performance Result

1. Improvement for query having join between TABLE_CONSTRAINTS and KEY_COLUMN_USAGE:

Objects	Before (ms)	After (ms)	Improvement %
1181 tables and primary key constraints	4781	87	98.18%
3000 tables and primary key constraints	33011	98	99.70%
3000 tables, primary key, unique and check constraints	14173	116	99.18%

2 Improvement for different queries:

Objects	Query	Before (ms)	After (ms)	Improvement %
1181 tables and primary key constraints	SELECT * FROM INFORMATION_SCHEMA.TABLE_CONSTRAINTS	29	32.2	-11.03%
^	SELECT * FROM INFORMATION_SCHEMA.TABLE_CONSTRAINTS WHERE TABLE_NAME = 't456'	38	33.6	11.58%
^	SELECT * from INFORMATION_SCHEMA.TABLE_CONSTRAINTS WHERE TABLE_SCHEMA = 'dbo'	13.8	16.8	-21.74%
^	SELECT * from INFORMATION_SCHEMA.TABLE_CONSTRAINTS WHERE CONSTRAINT_TYPE = 'PRIMARY KEY'	24.4	16.6	31.97%
^	SELECT * from INFORMATION_SCHEMA.KEY_COLUMN_USAGE	22	20.2	8.18%
^	SELECT * from INFORMATION_SCHEMA.KEY_COLUMN_USAGE WHERE TABLE_NAME = 't456'	8	6.4	20%
^	SELECT * from INFORMATION_SCHEMA.KEY_COLUMN_USAGE WHERE TABLE_SCHEMA = 'dbo'	21.4	19.2	10.28%
3000 tables, primary key, unique and check constraints	SELECT * FROM INFORMATION_SCHEMA.TABLE_CONSTRAINTS	105.2	128	-21.67%
^	SELECT * FROM INFORMATION_SCHEMA.TABLE_CONSTRAINTS WHERE TABLE_NAME = 't456'	79.8	58.2	27.07%
^	SELECT * from INFORMATION_SCHEMA.TABLE_CONSTRAINTS WHERE TABLE_SCHEMA = 'dbo'	77.8	100.6	-29.31%
^	SELECT * from INFORMATION_SCHEMA.TABLE_CONSTRAINTS WHERE CONSTRAINT_TYPE = 'PRIMARY KEY'	61.4	53.2	13.36%
^	SELECT * from INFORMATION_SCHEMA.KEY_COLUMN_USAGE	71.4	65.6	8.12%
^	SELECT * from INFORMATION_SCHEMA.KEY_COLUMN_USAGE WHERE TABLE_NAME = 't456'	12	10	16.67%
^	SELECT * from INFORMATION_SCHEMA.KEY_COLUMN_USAGE WHERE TABLE_SCHEMA = 'dbo'	75.6	67.2	11.11%

NOTE: We don't observe ANY degradation if we have 100 user schemas

Check List

• Commits are signed per the DCO using --signoff

By submitting this pull request, I confirm that my contribution is under the terms of the Apache 2.0 and PostgreSQL licenses, and grant any person obtaining a copy of the contribution permission to relicense all or a portion of my contribution to the PostgreSQL License solely to contribute all or a portion of my contribution to the PostgreSQL open source project.

For more information on following Developer Certificate of Origin and signing off your commits, please check here.

[coveralls]

Pull Request Test Coverage Report for Build 12314800454

Details

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 73.811%

---

Totals
Change from base Build 12314489542:	0.0%
Covered Lines:	43164
Relevant Lines:	58479

---

💛 - Coveralls