[CALCITE-2450] RexSimplify: reorder predicates to a canonical form as a part of RexSimplify

• Type: Sub-task
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.17.0
• Fix Version/s: 1.22.0
• Component/s: core
• Labels: pull-request-available
• Assignee: Vladimir Sitnikov
• Reporter: Vladimir Sitnikov
• Created: 07/Aug/18 09:19
• Updated: 07/Oct/22 23:52
• Resolved: 10/Jan/20 08:37

Description

Certain optimizations are easier to perform when input expressions are in a canonical form.
For instance: more duplicates can be found in AND/OR lists, case branches, etc.

Note: this reordering is supposed to happen in RexSimplify only. In other words, RexBuilder would still produce "non-canonical" expressions.

It is expected that RexSimplify might alter the expression, so if it converts 5=x to x=5 it should be just fine.

The suggested rules are to be discussed, yet the following might be fine:
1) For AND, OR, IN: put "simpler" nodes first. The weight of a node could be either .toString().length() or a number of child nodes or something like that.
The motivation is to simplify logic that handles "duplicate" entries. It won't have to consider "both alternatives" all over the place.
2) For comparison with literals put literal as the second argument
3) For binary comparison, put node with less weight to the left

Issue Links

• is related to: HIVE-26238 Decouple sort filter predicates optimization from digest normalization in CBO
• is related to: HIVE-26312 Use default digest normalization strategy in CBO
• relates to: CALCITE-5315 Error when pushing filters with operations that may throw (CAST/DIVISION) below joins
• relates to: CALCITE-3746 RexSimplify changes the order of IS NOT NULL in And RexNode
• links to: GitHub Pull Request #1703