feat(frontend): add delta join solver

[skyzh]

Signed-off-by: Alex Chi iskyzh@gmail.com

What's changed and what's your intention?

The solver generates the lookup plans by controlling two parameters StreamStrategy::LeftThisEpoch or ArrangeStrategy::LeftFirst and based on join order generated by optimizer.

Checklist

• I have written necessary docs and comments
• I have added necessary unit tests and integration tests

Refer to a related PR or issue link (optional)

ref #1962
close #2147

[codecov]

Codecov Report

Merging #2144 (25a1e06) into main (a315894) will increase coverage by 0.09%.
The diff coverage is 95.16%.

@@            Coverage Diff             @@
##             main    #2144      +/-   ##
==========================================
+ Coverage   70.51%   70.60%   +0.09%     
==========================================
  Files         645      649       +4     
  Lines       81785    82164     +379     
==========================================
+ Hits        57668    58010     +342     
- Misses      24117    24154      +37     

Flag	Coverage Δ
rust	70.60% <95.16%> (+0.09%)	⬆️

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

Impacted Files	Coverage Δ
src/frontend/src/optimizer/mod.rs	94.35% <ø> (ø)
src/frontend/src/optimizer/join_solver.rs	95.16% <95.16%> (ø)
src/stream/src/executor/barrier_align.rs	0.00% <0.00%> (-93.11%)	⬇️
src/stream/src/executor/mod.rs	43.14% <0.00%> (-5.43%)	⬇️
src/stream/src/executor_v2/debug/update_check.rs	92.85% <0.00%> (-0.90%)	⬇️
src/stream/src/executor_v2/debug/schema_check.rs	94.11% <0.00%> (-0.23%)	⬇️
src/meta/src/stream/stream_manager.rs	63.53% <0.00%> (-0.10%)	⬇️
src/stream/src/executor_v2/global_simple_agg.rs	94.71% <0.00%> (-0.08%)	⬇️
src/stream/src/executor_v2/hash_agg.rs	96.81% <0.00%> (-0.05%)	⬇️
src/stream/src/executor_v2/local_simple_agg.rs	97.60% <0.00%> (-0.05%)	⬇️
... and 26 more

📣 Codecov can now indicate which changes are the most critical in Pull Requests. Learn more

[fuyufjh]

LSTM

[st1page]

generally LGTM. if we can rename the join_solver as delta_join_solver?

[skyzh]

generally LGTM. if we can rename the join_solver as delta_join_solver?

+1. I've got a lot of rename requests recently, and I'd like to do this in one PR later.

[skyzh]

Just come up with a new arrange strategy: tables with larger join order should be placed to lookup prev epoch, while placing to right-most position 🤣

[yuhao-su]

LGTM. But for a, b, c on a.x=b.x and a.y=c.y case, we might another solver.

[skyzh]

Fixed several typos in this PR, and made arrange strategies to be left-most by default. After some discussion with @st1page, we decide to minimize the data on stream in all cases.