feat(optimizer): support agg group by simplify rule

[chenzl25]

I hereby agree to the terms of the RisingWave Labs, Inc. Contributor License Agreement.

What's changed and what's your intention?

• Resolve feat: Use functional dependencies to simplify aggregation's group by #9578
• Use first_value to replace redundant group by key, so that first_value could be purged by the following column pruning optimization.

Example:

before:
select count(*) from sbtest1 group by id, k, c;

use functional dependencies  (id) ==> (k, c)

after:
select count(*) from sbtest1 group by id;

Checklist

• I have written necessary rustdoc comments
• I have added necessary unit tests and integration tests
• I have added fuzzing tests or opened an issue to track them. (Optional, recommended for new SQL features Sqlsmith: Sql feature generation #7934).
• My PR contains breaking changes. (If it deprecates some features, please create a tracking issue to remove them in the future).
• All checks passed in ./risedev check (or alias, ./risedev c)
• My PR changes performance-critical code. (Please run macro/micro-benchmarks and show the results.)

• My PR contains critical fixes that are necessary to be merged into the latest release. (Please check out the details)

Documentation

• My PR needs documentation updates. (Please use the Release note section below to summarize the impact on users)

Release note

If this PR includes changes that directly affect users or other significant modifications relevant to the community, kindly draft a release note to provide a concise summary of these changes. Please prioritize highlighting the impact these changes will have on users.

[codecov]

Codecov Report

Merging #12349 (d973896) into main (7baa27f) will increase coverage by 0.00%.
Report is 7 commits behind head on main.
The diff coverage is 98.33%.

@@           Coverage Diff           @@
##             main   #12349   +/-   ##
=======================================
  Coverage   69.90%   69.91%           
=======================================
  Files        1415     1416    +1     
  Lines      235524   235583   +59     
=======================================
+ Hits       164644   164708   +64     
+ Misses      70880    70875    -5     

Flag	Coverage Δ
rust	69.91% <98.33%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files Changed	Coverage Δ
src/frontend/src/optimizer/rule/mod.rs	100.00% <ø> (ø)
...d/src/optimizer/rule/agg_group_by_simplify_rule.rs	98.14% <98.14%> (ø)
src/frontend/src/optimizer/logical_optimization.rs	97.96% <100.00%> (+0.01%)	⬆️

... and 4 files with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[st1page]

Furthermore, will we support that?

CREATE TABLE t(v int, k int primary key);
SELECT v, k from t group by k;
ERROR:  QueryError: Invalid input syntax: column must appear in the GROUP BY clause or be used in an aggregate function

[chenzl25]

Furthermore, will we support that?

CREATE TABLE t(v int, k int primary key);
SELECT v, k from t group by k;
ERROR:  QueryError: Invalid input syntax: column must appear in the GROUP BY clause or be used in an aggregate function

To be compatible with PostgreSQL, I think we don't need to support it, although other databases like MySQL support this feature.

[kwannoel]

Use first_value to replace redundant group by key, so that first_value could be purged by the following column pruning optimization.

How does column pruning purge the first_value. Seems like it is still present in some of the output plans?

[stdrc]

Maybe we need a any_value (exists in DuckDB while its semantic is first non-null value) or arbitrary_value instead of first_value. Currently streaming impl of first_value requires an order and thus a state table...

[chenzl25]

Use first_value to replace redundant group by key, so that first_value could be purged by the following column pruning optimization.

How does column pruning purge the first_value. Seems like it is still present in some of the output plans?

If those columns haven't been used, they could be pruned, however, if they are still being used, they will present.

[chenzl25]

Maybe we need a any_value (exists in DuckDB while its semantic is first non-null value) or arbitrary_value instead of first_value. Currently streaming impl of first_value requires an order and thus a state table...

Yes, but for each group key, this rule can ensure that there is always one row in the first_value state table corresponding to it.

[stdrc]

Maybe we need a any_value (exists in DuckDB while its semantic is first non-null value) or arbitrary_value instead of first_value. Currently streaming impl of first_value requires an order and thus a state table...

Yes, but for each group key, this rule can ensure that there is always one row in the first_value state table corresponding to it.

Aren't there #input rows of each group rows of the same value in each group?

[st1page]

Yes, but for each group key, this rule can ensure that there is always one row in the first_value state table corresponding to it.

Exactly not 🥵 there are multiple records in each group

CREATE TABLE User{
    id int Primary key,
    name varchar,    
}

CREATE TABLE Events {
    user_id int,
}

SELECT user_id, name, count(*) FROM User 
Join Event On Events.user_id = User.id
Group by user_id, name;

Btw, this case need us do agg-join push down rule 🤔