plan: fix insert onduplicate key update

[zz-jason]

What problem does this PR solve?

The schema we built for Insert.Schema4OnDuplicate was incorrect, thus the data record inserted by of insert ... select ... on duplicate key update ... statement based on this schema was incorrect, too. Let's take the following queries as an example:

drop table if exists t;
create table t(k1 bigint, k2 bigint, val bigint, primary key(k1, k2));
insert into t (val, k1, k2) values (3, 1, 2);
insert into t (val, k1, k2) select c, a, b from (select 1 as a, 2 as b, 4 as c) tmp on duplicate key update val = tmp.c;

After executing the above queries, the data record inside in table t is not as expected:

TiDB(localhost:4000) > select * from t;
+----+----+------+
| k1 | k2 | val  |
+----+----+------+
|  1 |  2 |    1 |
+----+----+------+
1 row in set (0.00 sec)

This PR fixes this issue, and the correct result is:

TiDB(localhost:4000) > select * from t;
+----+----+------+
| k1 | k2 | val  |
+----+----+------+
|  1 |  2 |    4 |
+----+----+------+
1 row in set (0.00 sec)

What is changed and how it works?

construct the Insert.Schema4OnDuplicate based on the schema of table to be inserted.

Check List

Tests

• Integration test

Related changes

• Need to be included in the release note

[zz-jason]

/run-all-tests

[zz-jason]

/run-all-tests

[alivxxx]

Is it affectedValuesCols?

[zz-jason]

Yeah, this name is better

[alivxxx]

And we already have it at line 1158.

[XuHuaiyu]

We need a comment for this variable.

[zz-jason]

@XuHuaiyu @lamxTyler PTAL

[XuHuaiyu]

LGTM

[alivxxx]

LGTM