executor: support GROUP_CONCAT(ORDER BY) (#16591)

[sre-bot]

cherry-pick #16591 to release-3.0

---

What problem does this PR solve?

Issue Number: close #6838

Problem Summary:

What is changed and how it works?

Proposal: xxx

What's Changed:

Executor part:

1. Add a heap in GROUP_CONCAT function implementation.
2. Build hash agg in unparalleled if group_concat has ORDER BY clause.

Planner part:

1. Build and resolve arguments of ORDER BY clause.
2. Add explain info with ORDER BY clause.
3. Handle arguments of ORDER BY clause in ResolveIndices, PruneColumns and InjectProjBelowAgg
4. Do not push down to coprocessor if group_concat has ORDER BY clause.
5. Do not push down to Union and Join if group_concat has ORDER BY clause.

How it Works:

mysql> desc select 1, 2, 3, 4, 5 , group_concat(name, id ORDER BY 2 desc, name SEPARATOR '++') from test;
+------------------------------+---------+-----------+---------------+------------------------------------------------------------------------------------------------------------------------------------------------------+
| id                           | estRows | task      | access object | operator info                                                                                                                                        |
+------------------------------+---------+-----------+---------------+------------------------------------------------------------------------------------------------------------------------------------------------------+
| Projection_4                 | 1.00    | root      |               | 1->Column#5, 2->Column#6, 3->Column#7, 4->Column#8, 5->Column#9, Column#4                                                                            |
| └─StreamAgg_8                | 1.00    | root      |               | funcs:group_concat(Column#15, Column#16 order by Column#17 desc, Column#18 asc separator "++")->Column#4                                             |
|   └─Projection_18            | 6.00    | root      |               | cast(test.test.name, var_string(20))->Column#15, cast(test.test.id, var_string(20))->Column#16, test.test.name->Column#17, test.test.name->Column#18 |
|     └─TableReader_15         | 6.00    | root      |               | data:TableFullScan_14                                                                                                                                |
|       └─TableFullScan_14     | 6.00    | cop[tikv] | table:test    | keep order:false, stats:pseudo                                                                                                                       |
+------------------------------+---------+-----------+---------------+------------------------------------------------------------------------------------------------------------------------------------------------------+
5 rows in set (0.00 sec)

mysql> select 1, 2, 3, 4, 5 , group_concat(name, id ORDER BY 2 desc, name SEPARATOR '++') from test;
+---+---+---+---+---+-------------------------------------------------------------+
| 1 | 2 | 3 | 4 | 5 | group_concat(name, id ORDER BY 2 desc, name SEPARATOR '++') |
+---+---+---+---+---+-------------------------------------------------------------+
| 1 | 2 | 3 | 4 | 5 | 5003++2003++301++201++202++101                              |
+---+---+---+---+---+-------------------------------------------------------------+
1 row in set (0.00 sec)

Related changes

• PR to update pingcap/docs/pingcap/docs-cn:
• Need to cherry-pick to the release branch

Check List

Tests

• Unit test
• Integration test

Side effects

• Performance regression
  • Consumes more MEM

Release note

• support ORDER BY clause in GROUP_CONCAT

[sre-bot]

/run-all-tests

[SunRunAway]

/run-all-tests

[SunRunAway]

/run-all-tests

[qw4990]

LGTM

[winoros]

lgtm

[SunRunAway]

/merge

[sre-bot]

/run-all-tests