Improve sharded query routing for tuple list

[harshit-gangal]

Description

This PR improved the plan when filter condition has List Argument.
E.g. select 1 from user where (id, col) in ::vals
Earlier this query was send down as Scatter to all shards.
With this PR, the query will send down to the shards based on where the sharding key (here id) routes this query.

These List Arguments is currently used in DML routing for multi-table query and foreign key cascades.

Old Plan:

{
  "QueryType": "SELECT",
  "Original": "select 1 from user where (id, col) in ::vals",
  "Instructions": {
    "OperatorType": "Route",
    "Variant": "Scatter",
    "Keyspace": {
      "Name": "user",
      "Sharded": true
    },
    "FieldQuery": "select 1 from `user` where 1 != 1",
    "Query": "select 1 from `user` where (id, col) in ::vals",
    "Table": "`user`"
  },
  "TablesUsed": [
    "user.user"
  ]
}

New Plan:

{
  "QueryType": "SELECT",
  "Original": "select 1 from user where (id, col) in ::vals",
  "Instructions": {
    "OperatorType": "Route",
    "Variant": "MultiEqual",
    "Keyspace": {
      "Name": "user",
      "Sharded": true
    },
    "FieldQuery": "select 1 from `user` where 1 != 1",
    "Query": "select 1 from `user` where (id, col) in ::vals",
    "Table": "`user`",
    "Values": [
      "vals:0"
    ],
    "Vindex": "user_index"
  },
  "TablesUsed": [
    "user.user"
  ]
}

Related Issue(s)

Checklist

• 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.

[codecov]

Codecov Report

Attention: 114 lines in your changes are missing coverage. Please review.

Comparison is base (eddb39e) 47.29% compared to head (825ce76) 47.28%.
Report is 3 commits behind head on main.

❗ Current head 825ce76 differs from pull request most recent head b4f7288. Consider uploading reports for the commit b4f7288 to get more accurate results

Files	Patch %	Lines
go/vt/vtgate/evalengine/expr_tuple_bvar.go	0.00%	50 Missing ⚠️
...vt/vtgate/planbuilder/operators/sharded_routing.go	0.00%	39 Missing ⚠️
go/vt/vtgate/evalengine/cached_size.go	0.00%	10 Missing ⚠️
go/vt/vtgate/evalengine/format.go	0.00%	6 Missing ⚠️
go/vt/vtgate/evalengine/translate_simplify.go	0.00%	4 Missing ⚠️
go/vt/vtgate/planbuilder/operators/route.go	0.00%	3 Missing ⚠️
go/vt/vtgate/evalengine/expr_bvar.go	0.00%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #14892      +/-   ##
==========================================
- Coverage   47.29%   47.28%   -0.02%     
==========================================
  Files        1137     1138       +1     
  Lines      238684   238787     +103     
==========================================
+ Hits       112895   112912      +17     
- Misses     117168   117257      +89     
+ Partials     8621     8618       -3     

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

[systay]

Suggested change

	return nil, vterrors.Errorf(vtrpcpb.Code_INVALID_ARGUMENT, "query argument '%s' must be a tuple (is %s)", bv.Key, bvar.Type)
	return nil, vterrors.Errorf(vtrpcpb.Code_INVALID_ARGUMENT, "query argument '%s' must be a tuple (is %s)", bv.Key, bvar.Type.String())

[harshit-gangal]

I can add that explicitly. Otherwise, It is not required when a String() method is implemented.

[dbussink]

Can we add the logic for this here? We need this also for the type resolving as well, since we depend on compile for resolving types now as well. See for example also here:

vitess/go/vt/vtgate/evalengine/translate.go

Lines 697 to 703 in c534201

	func (u *UntypedExpr) Compile(env *ExpressionEnv) (*CompiledExpr, error) {
	typed, err := u.loadTypedExpression(env)
	if err != nil {
	return nil, err
	}
	return typed.compile(u.ir, u.collation, u.collationEnv, env.sqlmode)
	}

[harshit-gangal]

This is a Tuple type bind variable and I am not sure how to compile such.
I checked expr_bvar and it is not recognized there

func (bvar *BindVariable) compile(c *compiler) (ctype, error) {
.
.
.
switch tt := typ.Type; {
.
.
default:
    return ctype{}, vterrors.Errorf(vtrpcpb.Code_UNIMPLEMENTED, "Type is not supported: %s", tt)
}
.		

[harshit-gangal]

We store something like
(1, 'a'), (2, 'b'), (3, 'c')... in a single bind variable.
and TupleBindVariable will provide single tuple for an index.
If Index is 1 then it returns
('a'), ('b'), ('c'). The output type is still a tuple.