planner: enhanced index range calculation plan

[xiaodong-ji]

What problem does this PR solve?

Issue Number: close #21280

Problem Summary: unexpected selection in index-look-up's build-side after add a probe-side condition.

What is changed and how it works?

What's Changed:

How it Works:

Related changes

Tests

• Manual test

CREATE TABLE `t` (
  `a` int(11) DEFAULT NULL,
  `b` int(11) DEFAULT NULL,
  `c` int(11) DEFAULT NULL,
  `d` int(11) DEFAULT NULL,
  KEY `kk` (`a`,`b`,`c`)
)
analyze table t;

mysql> explain select * from t where a = 1 and ((b = 1 and c = 1) or (b = 3)) and d = 3;
+-------------------------------+---------+-----------+----------------------------+----------------------------------------------------------------+
| id                            | estRows | task      | access object              | operator info                                                  |
+-------------------------------+---------+-----------+----------------------------+----------------------------------------------------------------+
| IndexLookUp_11                | 0.00    | root      |                            |                                                                |
| ├─IndexRangeScan_8(Build)     | 0.00    | cop[tikv] | table:t, index:kk(a, b, c) | range:[1 1 1,1 1 1], [1 3,1 3], keep order:false, stats:pseudo |
| └─Selection_10(Probe)         | 0.00    | cop[tikv] |                            | eq(test.t.d, 3)                                                |
|   └─TableRowIDScan_9          | 0.00    | cop[tikv] | table:t                    | keep order:false, stats:pseudo                                 |
+-------------------------------+---------+-----------+----------------------------+----------------------------------------------------------------+
4 rows in set (0.00 sec)

CREATE TABLE `tt` (
  `a` int(11) DEFAULT NULL,
  `b` int(11) DEFAULT NULL,
  `c` int(11) DEFAULT NULL,
  `d` int(11) DEFAULT NULL,
  KEY `kk` (`a`,`b`,`c`,`d`)
)

mysql> explain select * from tt where a = 1 and ((b = 1 and c = 1) or (b = 3)) and d = 3; 
+--------------------------+---------+-----------+--------------------------------+--------------------------------------------------------------------+
| id                       | estRows | task      | access object                  | operator info                                                      |
+--------------------------+---------+-----------+--------------------------------+--------------------------------------------------------------------+
| IndexReader_7            | 0.00    | root      |                                | index:Selection_6                                                  |
| └─Selection_6            | 0.00    | cop[tikv] |                                | eq(test.tt.d, 3)                                                   |
|   └─IndexRangeScan_5     | 0.00    | cop[tikv] | table:tt, index:kk(a, b, c, d) | range:[1 1 1 3,1 1 1 3], [1 3,1 3], keep order:false, stats:pseudo |
+--------------------------+---------+-----------+--------------------------------+--------------------------------------------------------------------+
3 rows in set (0.00 sec)

Release note

• enhancement index range calculation plan

[Reminiscent]

It's better to add some comments to explain the logic of the code. Thanks!

[Reminiscent]

/run-all-tests

[Reminiscent]

/run-all-tests

[xiaodong-ji]

/run-all-tests

[xiaodong-ji]

/run-all-tests

[xiaodong-ji]

/run-all-tests

[ti-chi-bot]

@xiaodong-ji: PR needs rebase.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.