Enhance data schema generation for empty response

[Jackie-Jiang]

This is a followup on #14836 with the following enhancements:

• Handle MV column type properly
• Handle the case when different number of columns are returned in single-stage vs multi-stage engine
• More exception handling to prevent it from throwing exception

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 0% with 39 lines in your changes missing coverage. Please review.

Project coverage is 63.73%. Comparing base (59551e4) to head (5c7ac6c).
Report is 1632 commits behind head on master.

Files with missing lines	Patch %	Lines
...g/apache/pinot/query/parser/utils/ParserUtils.java	0.00%	37 Missing ⚠️
...sthandler/BaseSingleStageBrokerRequestHandler.java	0.00%	2 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff              @@
##             master   #14918      +/-   ##
============================================
+ Coverage     61.75%   63.73%   +1.98%     
- Complexity      207     1471    +1264     
============================================
  Files          2436     2708     +272     
  Lines        133233   151539   +18306     
  Branches      20636    23398    +2762     
============================================
+ Hits          82274    96580   +14306     
- Misses        44911    47704    +2793     
- Partials       6048     7255    +1207     

Flag	Coverage Δ
custom-integration1	100.00% <ø> (+99.99%)	⬆️
integration	100.00% <ø> (+99.99%)	⬆️
integration1	100.00% <ø> (+99.99%)	⬆️
integration2	0.00% <ø> (ø)
java-11	63.71% <0.00%> (+2.00%)	⬆️
java-21	63.62% <0.00%> (+2.00%)	⬆️
skip-bytebuffers-false	63.72% <0.00%> (+1.98%)	⬆️
skip-bytebuffers-true	63.58% <0.00%> (+35.86%)	⬆️
temurin	63.73% <0.00%> (+1.98%)	⬆️
unittests	63.72% <0.00%> (+1.98%)	⬆️
unittests1	56.31% <0.00%> (+9.42%)	⬆️
unittests2	34.01% <0.00%> (+6.27%)	⬆️

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

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

[albertobastos]

Thanks for the enhancements Jackie, looks good to me.

[gortiz]

I have contradictory feelings about explaining why some issues can happen in the generic method that solves them. On the one hand, it is useful to have the reasons centralized. On the other hand, it is a clear candidate for documentation that could be desynchronized with the code (ie we change one of these cases and forget to change the javadoc).

[Jackie-Jiang]

We probably need a dev focus documentation to document the internal behaviors. It is always a hard problem to keep the doc in sync with code though. I'd prefer having it in javadoc for now for future developers to know the history