Fix multicolumn parquet predicate pushdown (#4046)

[tustvold]

Which issue does this PR close?

Closes #4046

Rationale for this change

The first incarnation of this in #4021 computed the mapping from the projected schema to the file schema and not the other way round. For the case of two columns these are always the same thing and so it worked for the two column case, but more columns requires additional shenanigans.

What changes are included in this PR?

Are there any user-facing changes?

[alamb]

Thank you @tustvold

This is all so familiar -- we have very similar logic in IOx in https://github.com/influxdata/influxdb_iox/blob/main/schema/src/sort.rs (though I don't see any obvious way to reuse it)

[alamb]

Thank you for these comments

[alamb]

I verified this fixes the issue in my test harness in #3976 -- given this is a non controversial bug fix I am merging it in now to keep that project chugging along

[ursabot]

Benchmark runs are scheduled for baseline = c92a38f and contender = ea31da9. ea31da9 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