pingcap/tidb

Support to use the outer side of the (Anti) Semi Join to build the hash table

Open

#32,335 opened on Feb 14, 2022

View on GitHub
 (9 comments) (4 reactions) (1 assignee)Go (40,090 stars) (6,186 forks)batch import
help wantedsig/plannertype/enhancement

Description

Enhancement

Currently, when TiDB meets a(an) (Anti) Semi join, it would always choose the inner side to build the hash table. When the inner side is extremely large compared with the outer side, our current strategy would cost a lot on building the hash table by the inner side.

So we should support constructing the hash table by the outer side to speed up the execution when the data distribution is like the above case.

It's something the same with supporting constructing hash table by the outer side of the outer join, which is implemented in TiDB.

Contributor guide

Tech stack
go
Domain
database
Issue type
feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
4
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
over 1 week
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
blocked
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
clear
Prerequisites
Go programminghash join algorithmTiDB execution engineSQL joins
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
15
Research direction
Investigate the existing implementation of hash table construction for outer joins in TiDB, located in the executor package (e.g., `executor/join.go`). Review the comments on this issue (#32335) for design discussions and potential pitfalls. Study how the semi/anti join logic works in the hash join operator and consider how to adapt the outer join approach to support building the hash table from the outer side when it is smaller.
Support to use the outer side of the (Anti) Semi Join to build the hash table · pingcap/tidb#32335 | Good First Issue