[WIP] executor: skip null rows when deleting multiple tables

[ekexium]

Signed-off-by: ekexium [email protected]

What problem does this PR solve?

Issue Number: close #31321

Problem Summary:

Skip null rows when deleting using outer joins.

I'm not sure whether it's an appropriate way to fix this. Please be careful when reviewing.

What is changed and how it works?

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

None

[ti-chi-bot]

[REVIEW NOTIFICATION]

This pull request has not been approved.

To complete the pull request process, please ask the reviewers in the list to review by filling /cc @reviewer in the comment.
After your PR has acquired the required number of LGTMs, you can assign this pull request to the committer in the list by filling /assign @committer in the comment to help you merge this pull request.

The full list of commands accepted by this bot can be found here.

Reviewer can indicate their review by submitting an approval review.
Reviewer can cancel approval by submitting a request changes review.

[sre-bot]

Code Coverage Details: https://codecov.io/github/pingcap/tidb/commit/d6d7b2b28afb5e72d52e95f3c62fb4a7e3d07371

[ekexium]

Only using null datums as a condition is bad. An example:

create table c(k1 int);
create table d(id int, k1 int);
insert into c(k1) values(null);
insert into d values(null, null);
delete c, d from c join d;
select * from d;

Before:

mysql> select * from d;
Empty set (0.00 sec)

This PR:

mysql> select * from d;
+------+------+
| id   | k1   |
+------+------+
| NULL | NULL |
+------+------+

[cfzjywxk]

Seems this will affect the result of delete c, d from c join d on c.k1 is null, or something like DELETE a FROM mt_order_delivery_address AS a LEFT JOIN mt_order AS o USING (order_id) WHERE o.order_id IS NULL;

Sometimes the null row is fetched from storage and sometimes they are generated by the left join operator. We may need to refer the mysql code to investigate the expected behaviour for such cases.

[ekexium]

It seems to me that MySQL uses an explicit variable to indicate null rows generated by outer joins.

set_null_row() and reset_null_row() are used by the join executor to
signal the presence or absence of a NULL-extended row for an outer joined
table.

https://github.com/mysql/mysql-server/blob/3290a66c89eb1625a7058e0ef732432b6952b435/sql/table.h#L1961-L1975

And there is a check to skip such rows when deleting.

https://github.com/mysql/mysql-server/blob/3290a66c89eb1625a7058e0ef732432b6952b435/sql/sql_delete.cc#L1040-L1041

[ti-chi-bot]

@ekexium: 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.