feat(optimizer): change stream join mv distribution key

[chenzl25]

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

What's changed and what's your intention?

• Change the stream join mv distribution key to be the same as the stream key to avoid slow backfilling caused by data skew of the dimension table join key.
• Resolve Discussion: use stream_key as MV's distribution key #12824

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.

[fuyufjh]

Sorry, can you take a look on #12824 (comment)

[chenzl25]

Sorry, can you take a look on #12824 (comment)

Currently, we just have stream_key. After we have the unique_key property we can change the distribution to it. It would be more elegant I think.

[fuyufjh]

Sorry, can you take a look on #12824 (comment)

Currently, we just have stream_key. After we have the unique_key property we can change the distribution to it. It would be more elegant I think.

Okay. Make sense to me.

But unique_key sounds ambiguous with the unique key in relation DB. Even though sharing some similarity, they seem to be different things to me.

[fuyufjh]

The code LGTM.

Needs to run a benchmark test before merging.

[chenzl25]

But unique_key sounds ambiguous with the unique key in relation DB. Even though sharing some similarity, they seem to be different things to me.

The unique_key should be exactly what it means in a relational Database. While stream_key is also a unique key, it might contain more redundant columns.

[codecov]

Codecov Report

Merging #13022 (117fdde) into main (e64543f) will increase coverage by 0.06%.
Report is 2 commits behind head on main.
The diff coverage is 91.66%.

@@            Coverage Diff             @@
##             main   #13022      +/-   ##
==========================================
+ Coverage   68.63%   68.70%   +0.06%     
==========================================
  Files        1495     1495              
  Lines      251135   251145      +10     
==========================================
+ Hits       172369   172548     +179     
+ Misses      78766    78597     -169     

Flag	Coverage Δ
rust	68.70% <91.66%> (+0.06%)	⬆️

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

Files	Coverage Δ
...tend/src/optimizer/plan_node/stream_materialize.rs	97.43% <100.00%> (+0.10%)	⬆️
...rc/frontend/src/optimizer/property/distribution.rs	76.23% <100.00%> (ø)
src/storage/src/hummock/file_cache/store.rs	40.49% <0.00%> (-0.17%)	⬇️

... and 20 files with indirect coverage changes

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