Jayanta Mondal d3c5128254 Merged PR 1405332: Mongo collation support Part 1 ( Find/Aggregate queries with runtime collation) ...

### Does this PR have any customer impact?
Yes

### Type (Feature, Refactoring, Bugfix, DevOps, Testing, Perf, etc)
Feature

### Does it involve schema level changes? (Table, Column, Index, UDF, etc level changes)
No

### Are you introducing any new config? If yes, do you have tests with and without them being set?
Yes (helio_api.enableCollation). Yes.

### Release Note (Refer [Template](../docs/release_notes/pgmongo_template.md))
Done

### Description

Runtime implementation of collation support in Mongo Find query.

We parse the Mongo collation spec and covert it to an ICU standard collation string. An example string is "und-u-ks-level1-kc-true  This string uniquely identify collation-based string comparison logic by Postgres.  See: https://www.postgresql.org/docs/current/collation.html

This document tracks the current support matrix for query operators

| Operator  | Runtime | Index Pushdown | Comments
| --------- | ---- | ----- | ----
|  $eq | Yes | No |
|  $gt | Yes |  No |
|  $gte | Yes | No |
|  $lt | Yes |  No |
|  $lte | Yes |  No |
|  $ne | Yes |  No |
|  $in | No* |  No | See below
|  $nin | No* |  No | Do
|  $and | Yes |  No |
|  $not | Yes |  No |
|  $nor | Yes |  No |
|  $or | Yes |  No |
|  $all | Yes |  No |
|  $elemMatch |  Yes   |  No |
|  $positional | N/A | N/A | positional query operators are only used in projections.
|  $regex | N/A |  N/A   | $regex ignores collations in Native Mongo
|  $exists | N/A  |   N/A   |
|  $size | N/A  |  N/A |
|  $type | N/A |   N/A |
|  $bitsAllClear | N/A  |  N/A  | Not relevant for Utf8
|  $bitsAnyClear | N/A  |  N/A  | Do
|  $bitsAllSet | N/A  |  N/A  | Do
|  $bitsAnySet |  N/A |  N/A  | Do
|  $mod | N/A  |  N/A  | Do
|  $range | N/A  |  N/A | Only applies to Number
|  $text | N/A  |  N/A  | Seprate Index, has its own collation spec
|  $geowithin | N/A   |  N/A  | Not relevant for Utf8
|  $geointersects | N/A    | N/A   | Do

* $in/$nin is implemented with HashMap that does not support collation friendly Hashkeys
* $regex : collations are ignored with
* $text: $text spec has collation specific parameters
* $rage: For numbers only

This document tracks the current support matrix for update operators/commands

| Operator  | Runtime | Index Pushdown |
| --------- | ---- | ----- |
|  find_and_modify | No | No |
|  updateOne | No |  No |
|  updateMany | No | No |
|  createIndex | No | No |

Tests:
1. All Gated Tests pass when Collation is enabled by default

TODO:
1. Index support [already implemented - but breaking down for easy review]
2. Persist collation info while collection created [already implemented - but breaking down for easy review]
3. $cond, $let

Related work items: #3412359

2024-08-26 13:12:13 +03:00

..

collation.h

Merged PR 1405332: Mongo collation support Part 1 ( Find/Aggregate queries with runtime collation)

2024-08-26 13:12:13 +03:00