This is an automated cherry-pick of #40593

What problem does this PR solve?

Issue Number: close #40535

Problem Summary:

The minimal reproduce step can be simplified to:

CREATE TABLE `t1` (
  `c1` bigint(20) NOT NULL DEFAULT '-2312745469307452950',
  `c2` datetime DEFAULT '5316-02-03 06:54:49',
  `c3` tinyblob DEFAULT NULL,
  PRIMARY KEY (`c1`) /*T![clustered_index] CLUSTERED */
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_bin;

CREATE TABLE `t2` (
  `c1` set('kn8pu','7et','vekx6','v3','liwrh','q14','1met','nnd5i','5o0','8cz','l') DEFAULT '7et,vekx6,liwrh,q14,1met',
  `c2` float DEFAULT '1.683167',
  KEY `k1` (`c2`,`c1`),
  KEY `k2` (`c2`)
) ENGINE=InnoDB DEFAULT CHARSET=gbk COLLATE=gbk_chinese_ci;

( select   /*+ agg_to_cop()  */  locate( `t1`.`c3` , `t1`.`c3` ) as r0 , `t1`.`c3` as r1 from `t1` where not( IsNull( `t1`.`c1` ) )   order by r0,r1  ) union all ( select     concat_ws(',', `t2`.`c2` , `t2`.`c1` ) as r0 , `t2`.`c1` as r1 from `t2`    order by r0,r1  ) order by 1 limit 273;

logical plan before logicalOptimize:

*core.LogicalLimit
  *core.LogicalSort
    *core.LogicalUnionAll
      *core.LogicalProjection
        *core.LogicalSort
          *core.LogicalProjection
            *core.DataSource
      *core.LogicalProjection
        *core.LogicalSort
          *core.LogicalProjection
            *core.DataSource   

logical plan after logicalOptimize:

*core.LogicalTopN
  *core.LogicalUnionAll
    *core.LogicalProjection
      *core.LogicalProjection
        *core.LogicalTopN
          *core.DataSource
    *core.LogicalProjection
      *core.LogicalProjection
        *core.LogicalTopN <-
          *core.DataSource

In planner/core/task.go:*PhysicalTopN.attach2Task： TopN.ByItems contains a column that cannot be obtained from copTask.indexPlan, so the optimizer chooses to use copTask.tablePlan, but copTask.tablePlan is nil, this results in "null pointer dereference".

What is changed and how it works?

There is a column in topN.ByItems is generated by proj, when topN is pushed down below proj, topN cannot obtain the column from datasource, to fix this, we can check whether topN.ByItems contains a column(with ID=0) generated by proj, if so, proj will prevent the optimizer from pushing topN down.

1. A column with ID=0 indicates that the column cannot be resolved by data source.
2. To determine whether a column is generated by projection, we can check whether the child of proj contains the column.

Check List

Tests

• Unit test
• Integration test
• Manual test (add detailed scripts or steps below)
• No code

Side effects

• Performance regression: Consumes more CPU
• Performance regression: Consumes more Memory
• Breaking backward compatibility

Documentation

• Affects user behaviors
• Contains syntax changes
• Contains variable changes
• Contains experimental features
• Changes MySQL compatibility

Release note

Please refer to Release Notes Language Style Guide to write a quality release note.

None