Add evalMv for MV predicate filtering (incl. REGEXP_LIKE)

[xiangfu0]

What does this PR change?

• Adds evalMv as both:
• a transform function (EvalMvTransformFunction) for block-based execution.
• a scalar function (EvalMvScalarFunction) for scalar execution paths.
• Registers evalMv in TransformFunctionType and TransformFunctionFactory.
• Adds MV predicate parsing/evaluation in EvalMvPredicateEvaluator.
• Supports predicate operators: =, !=, >, >=, <, <=, IN, NOT IN, BETWEEN, LIKE, REGEXP_LIKE, plus AND / OR / NOT.
• Adds/extends tests:
• EvalMvTransformFunctionTest unit coverage for REGEXP_LIKE.
• ArrayTest integration coverage for REGEXP_LIKE with useBothQueryEngines.

evalMv(mv_col, 'predicate') evaluates each MV element against the predicate and returns a filtered MV array.

Why?

This addresses MV filtering semantics gaps (notably with MSE + MV columns) where filtering can keep all MV elements from a row once any element matches. evalMv provides explicit per-element filtering without requiring CROSS JOIN UNNEST.

Query examples

1) MSE: filter MV values by LIKE

SELECT DISTINCT arrayToMv(evalMv(norm_url_list_combined, 'norm_url_list_combined LIKE ''%api%''')) AS sf_node_name
FROM rum_otel_data_p
WHERE product IN ('WEB')
  AND org_id = 'AAAAAAAAAAA'
  AND start_time_micros >= 1769715945540000
  AND norm_url_list_combined IS NOT NULL
ORDER BY sf_node_name ASC
LIMIT 50
OPTION(useMultistageEngine=true)

2) Both engines: filter MV values by REGEXP_LIKE

SELECT evalMv(stringArrayCol, 'REGEXP_LIKE(stringArrayCol, ''^/api/.*'')')
FROM ArrayTest
LIMIT 1

Expected first-row output: ['/api/v1', '/api/v2'] for source ['/api/v1', '/home', '/api/v2', '/metrics'].

3) Both engines: use filtered MV result in row predicate

SELECT COUNT(*)
FROM ArrayTest
WHERE arrayLength(evalMv(stringArrayCol, 'REGEXP_LIKE(stringArrayCol, ''^/api/.*'')')) = 2

Validation

• ./mvnw -pl pinot-core -am -Dtest=EvalMvTransformFunctionTest -Dsurefire.failIfNoSpecifiedTests=false test
• ./mvnw -pl pinot-integration-tests -am -Dtest=ArrayTest -Dsurefire.failIfNoSpecifiedTests=false test

[codecov-commenter]

Codecov Report

❌ Patch coverage is 22.02729% with 400 lines in your changes missing coverage. Please review.
✅ Project coverage is 63.12%. Comparing base (7305eec) to head (44db7aa).
⚠️ Report is 21 commits behind head on master.

Files with missing lines	Patch %	Lines
...r/transform/function/EvalMvPredicateEvaluator.java	22.76%	167 Missing and 6 partials ⚠️
...or/transform/function/EvalMvTransformFunction.java	25.43%	118 Missing and 11 partials ⚠️
...not/core/function/scalar/EvalMvScalarFunction.java	14.03%	98 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff              @@
##             master   #17659      +/-   ##
============================================
- Coverage     63.16%   63.12%   -0.05%     
- Complexity     1479     1500      +21     
============================================
  Files          3173     3177       +4     
  Lines        189917   190836     +919     
  Branches      29064    29216     +152     
============================================
+ Hits         119970   120464     +494     
- Misses        60621    61024     +403     
- Partials       9326     9348      +22     

Flag	Coverage Δ
custom-integration1	100.00% <ø> (ø)
integration	100.00% <ø> (ø)
integration1	100.00% <ø> (ø)
integration2	0.00% <ø> (ø)
java-11	63.09% <22.02%> (-0.05%)	⬇️
java-21	63.08% <22.02%> (-0.04%)	⬇️
temurin	63.12% <22.02%> (-0.05%)	⬇️
unittests	63.12% <22.02%> (-0.05%)	⬇️
unittests1	55.48% <22.02%> (-0.05%)	⬇️
unittests2	33.98% <3.31%> (-0.05%)	⬇️

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.

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.