use_cascades hint works incorrectly on partition table

[Illyrix]

Description

Bug Report

Please answer these questions before submitting your issue. Thanks!

1. What did you do?

execute:

create table t(id int not null auto_increment, col_int int not null, col_float float, primary key(id, col_int)) partition by range(col_int) (partition p0 values less than (100), partition pn values less than (MAXVALUE));
insert into t values(1, 10, 1), (101, 100, 101);

SELECT /*+ use_cascades(TRUE)*/ * from t;

2. What did you expect to see?

same result as normal select statement:

SELECT * from t;
+-----+---------+-----------+
| id  | col_int | col_float |
+-----+---------+-----------+
| 101 |     100 |       101 |
|   1 |      10 |         1 |
+-----+---------+-----------+
2 rows in set (0.00 sec)

3. What did you see instead?

SELECT /*+ use_cascades(TRUE)*/ * from t;
Empty set (0.00 sec)

4. What version of TiDB are you using? (tidb-server -V or run select tidb_version(); on TiDB)

5.7.25-TiDB-v4.0.0-beta.2-305-g94011e606

SIG slack channel

#sig-planner

Score

• 300

Mentor

• @lzmhhh123

[siddontang]

Great catch!!!

[francis0407]

This is because cascades planner is still under development. And partition table hasn't been supported yet.

[Enochack]

/assign