Fix logical plan serialization

[thinkharderdev]

Which issue does this PR close?

Closes #3555

Rationale for this change

Currently round-trip logical plan serialization is broken for some aggregation queries. This happens when the TypeCoercion optimizer casts literal values in aggregate function expressions and then the const evaluator optimizes out the cast (replacing with a literal for the casted value). This effectively changes the expression name and causes a mismatch when rebuilding the plan after deserialization.

What changes are included in this PR?

In TypeCoercion special-case the handling of aggregation function expressions to make sure we alias back to the original expression name if it has changed. This should ensure the schema is still valid and the plan can be rebuilt from its serialized representation.

Are there any user-facing changes?

no

no

[thinkharderdev]

@alamb @andygrove This feels like a bit of hacky special casing but seemed like the least insrusive way to deal with this. If optimizers can change expression names in general though this could come up in other places w/r/t serde of aggregation plans where we rely explicitly on the expression name.

[alamb]

I agree this seems a little icky (like it is treating the symptom rather than the root cause) but I think it is tested and seems reasonable to me

[alamb]

I wish I had a better story for you

[alamb]

I believe this PR is going to fail its checks until is it merged to include #3576

[ursabot]

Benchmark runs are scheduled for baseline = 634d912 and contender = 488be64. 488be64 is a master commit associated with this PR. Results will be available as each benchmark for each run completes.
Conbench compare runs links:
[Skipped ⚠️ Benchmarking of arrow-datafusion-commits is not supported on ec2-t3-xlarge-us-east-2] ec2-t3-xlarge-us-east-2
[Skipped ⚠️ Benchmarking of arrow-datafusion-commits is not supported on test-mac-arm] test-mac-arm
[Skipped ⚠️ Benchmarking of arrow-datafusion-commits is not supported on ursa-i9-9960x] ursa-i9-9960x
[Skipped ⚠️ Benchmarking of arrow-datafusion-commits is not supported on ursa-thinkcentre-m75q] ursa-thinkcentre-m75q
Buildkite builds:
Supported benchmarks:
ec2-t3-xlarge-us-east-2: Supported benchmark langs: Python, R. Runs only benchmarks with cloud = True
test-mac-arm: Supported benchmark langs: C++, Python, R
ursa-i9-9960x: Supported benchmark langs: Python, R, JavaScript
ursa-thinkcentre-m75q: Supported benchmark langs: C++, Java

[alamb]

I wonder if #3568 could be related

[thinkharderdev]

I wonder if #3568 could be related

Yeah, if I understand that ticket correctly then it's the same underlying issue