Make sure derived table column names are handled correctly

[systay]

Description

For some situations, during planning we create derived tables. When this happens, we sometimes get the column names wrong, which leads to queries that are not correct.

Example query, with bad plan:

select id, id from user 
union 
select id, bar from user_extra 
order by 1

bad plan produced:

{
  "QueryType": "SELECT",
  "Original": "select id, id from user union select id, bar from user_extra order by 1",
  "Instructions": {
    "OperatorType": "Sort",
    "Variant": "Memory",
    "OrderBy": "(2|3) ASC",
    "ResultColumns": 2,
    "Inputs": [
      {
        "OperatorType": "Distinct",
        "Collations": [
          "(0:4)",
          "(1:4)",
          "(2:3)",
          "3"
        ],
        "Inputs": [
          {
            "OperatorType": "Route",
            "Variant": "Scatter",
            "Keyspace": {
              "Name": "user",
              "Sharded": true
            },
            "FieldQuery": "select dt.id, dt.id, id, weight_string(id), weight_string(dt.id) from (select id, id from `user` where 1 != 1 union select id, bar from user_extra where 1 != 1) as dt where 1 != 1",
            "Query": "select dt.id, dt.id, id, weight_string(id), weight_string(dt.id) from (select id, id from `user` union select id, bar from user_extra) as dt",
            "Table": "`user`, user_extra"
          }
        ]
      }
    ]
  },
  "TablesUsed": [
    "user.user",
    "user.user_extra"
  ]
}

As you can see from the plan, one of the queries being sent down to mysql looks like:

select dt.id, dt.id, id, weight_string(id), weight_string(dt.id)
from (
	select id, id from `user` 
	union 
	select id, bar from user_extra) as dt

This query would fail in MySQL with an ambiguous column error for dt.id.

Since this change is quite involved, and these queries have never worked well, I suggest we don't backport this change set.

Related Issue(s)

Checklist

• "Backport to:" labels have been added if this change should be back-ported to release branches
• If this change is to be back-ported to previous releases, a justification is included in the PR description
• Tests were added or are not required
• Did the new or modified tests pass consistently locally and on CI?
• Documentation was added or is not required

Deployment Notes

[vitess-bot]

Review Checklist

Hello reviewers! 👋 Please follow this checklist when reviewing this Pull Request.

General

• Ensure that the Pull Request has a descriptive title.
• Ensure there is a link to an issue (except for internal cleanup and flaky test fixes), new features should have an RFC that documents use cases and test cases.

Tests

• Bug fixes should have at least one unit or end-to-end test, enhancement and new features should have a sufficient number of tests.

Documentation

• Apply the release notes (needs details) label if users need to know about this change.
• New features should be documented.
• There should be some code comments as to why things are implemented the way they are.
• There should be a comment at the top of each new or modified test to explain what the test does.

New flags

• Is this flag really necessary?
• Flag names must be clear and intuitive, use dashes (-), and have a clear help text.

If a workflow is added or modified:

• Each item in Jobs should be named in order to mark it as required.
• If the workflow needs to be marked as required, the maintainer team must be notified.

Backward compatibility

• Protobuf changes should be wire-compatible.
• Changes to _vt tables and RPCs need to be backward compatible.
• RPC changes should be compatible with vitess-operator
• If a flag is removed, then it should also be removed from vitess-operator and arewefastyet, if used there.
• vtctl command output order should be stable and awk-able.

[GrahamCampbell]

Is this aiming for the 19.0.2 release?

[codecov]

Codecov Report

Attention: Patch coverage is 70.58824% with 70 lines in your changes are missing coverage. Please review.

Project coverage is 68.41%. Comparing base (ca2659d) to head (340956c).

Files	Patch %	Lines
go/vt/vtgate/planbuilder/operators/aggregator.go	25.00%	30 Missing ⚠️
go/vt/vtgate/planbuilder/operators/projection.go	63.15%	7 Missing ⚠️
go/vt/vtgate/planbuilder/operators/route.go	89.36%	5 Missing ⚠️
...vtgate/planbuilder/operators/projection_pushing.go	88.00%	3 Missing ⚠️
go/vt/vtgate/planbuilder/operators/apply_join.go	94.73%	2 Missing ⚠️
go/vt/vtgate/planbuilder/operators/comments.go	0.00%	2 Missing ⚠️
go/vt/vtgate/planbuilder/operators/distinct.go	50.00%	2 Missing ⚠️
go/vt/vtgate/planbuilder/operators/filter.go	0.00%	2 Missing ⚠️
go/vt/vtgate/planbuilder/operators/hash_join.go	92.00%	2 Missing ⚠️
go/vt/vtgate/planbuilder/operators/horizon.go	0.00%	2 Missing ⚠️
... and 7 more

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #15588      +/-   ##
==========================================
- Coverage   68.43%   68.41%   -0.03%     
==========================================
  Files        1558     1558              
  Lines      196186   196361     +175     
==========================================
+ Hits       134269   134331      +62     
- Misses      61917    62030     +113     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[GuptaManan100]

Rest looks good to me!

[harshit-gangal]

nit: more comments can be added here as adding nil to semtable tables is not clear. I understand the intention here but comments will help.

[harshit-gangal]

Ideally we can add the DerivedTable here which is created below.

[systay]

the Derived table that we create here is of a very different type than what we store in the semantic state, and the info that we normally store there does not really make sense to store here. we just need an ID

[harshit-gangal]

Looks good overall. some nits

[deepthi]

@systay is there a user-reported issue for this fix that is still open and should be closed?

[systay]

@systay is there a user-reported issue for this fix that is still open and should be closed?

no, this was found during our own testing