Chore: Use DataFusion impl of date_trunc function

[kazantsev-maksim]

Which issue does this PR close?

Rationale for this change

What changes are included in this PR?

How are these changes tested?

[andygrove]

Thanks @kazantsev-maksim

[comphead]

There are some compilation issues

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 41.39%. Comparing base (f09f8af) to head (e17e4ae).
⚠️ Report is 577 commits behind head on main.

Additional details and impacted files

@@              Coverage Diff              @@
##               main    #2523       +/-   ##
=============================================
- Coverage     56.12%   41.39%   -14.74%     
- Complexity      976     1068       +92     
=============================================
  Files           119      146       +27     
  Lines         11743    13562     +1819     
  Branches       2251     2358      +107     
=============================================
- Hits           6591     5614      -977     
- Misses         4012     6975     +2963     
+ Partials       1140      973      -167     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[parthchandra]

I seem to recall that we implemented this because the DF version was slower.

[comphead]

This is a good point @parthchandra we can "steal" date_trunc bench from DF and check it out.

[andygrove]

Current Comet performance for date_trunc isn't great according to CometDatetimeExpressionBenchmark.

Running benchmark: Date Truncate  - YEAR
  Running case: Date Truncate  - YEAR - Spark 
  Stopped after 38 iterations, 2006 ms
  Running case: Date Truncate  - YEAR - Comet
  Stopped after 39 iterations, 2022 ms

OpenJDK 64-Bit Server VM 11.0.28+6-post-Ubuntu-1ubuntu122.04.1 on Linux 6.8.0-84-generic
AMD Ryzen 9 7950X3D 16-Core Processor
Date Truncate  - YEAR:                    Best Time(ms)   Avg Time(ms)   Stdev(ms)    Rate(M/s)   Per Row(ns)   Relative
------------------------------------------------------------------------------------------------------------------------
Date Truncate  - YEAR - Spark                        46             53           6         23.0          43.5       1.0X
Date Truncate  - YEAR - Comet                        46             52           5         22.6          44.2       1.0X

Running benchmark: Date Truncate  - YYYY
  Running case: Date Truncate  - YYYY - Spark 
  Stopped after 46 iterations, 2019 ms
  Running case: Date Truncate  - YYYY - Comet
  Stopped after 42 iterations, 2006 ms

OpenJDK 64-Bit Server VM 11.0.28+6-post-Ubuntu-1ubuntu122.04.1 on Linux 6.8.0-84-generic
AMD Ryzen 9 7950X3D 16-Core Processor
Date Truncate  - YYYY:                    Best Time(ms)   Avg Time(ms)   Stdev(ms)    Rate(M/s)   Per Row(ns)   Relative
------------------------------------------------------------------------------------------------------------------------
Date Truncate  - YYYY - Spark                        39             44           6         26.6          37.6       1.0X
Date Truncate  - YYYY - Comet                        43             48           6         24.4          40.9       0.9X

Running benchmark: Date Truncate  - YY
  Running case: Date Truncate  - YY - Spark 
  Stopped after 51 iterations, 2010 ms
  Running case: Date Truncate  - YY - Comet
  Stopped after 44 iterations, 2040 ms

OpenJDK 64-Bit Server VM 11.0.28+6-post-Ubuntu-1ubuntu122.04.1 on Linux 6.8.0-84-generic
AMD Ryzen 9 7950X3D 16-Core Processor
Date Truncate  - YY:                      Best Time(ms)   Avg Time(ms)   Stdev(ms)    Rate(M/s)   Per Row(ns)   Relative
------------------------------------------------------------------------------------------------------------------------
Date Truncate  - YY - Spark                          37             39           3         28.6          34.9       1.0X
Date Truncate  - YY - Comet                          43             46           6         24.5          40.8       0.9X

I tried running against this PR but got similar failures to the ones we are seeing in CI.

[kazantsev-maksim]

The DataFusion implementation of date_trunc has a narrower scope than Spark's, as it only accepts TimestampType, while Spark's version additionally supports StringType and DateType.

Comet does not fully support casting DateType to TimestampType yet.