ci(query): cover cross-type nullable correlated joins

[youngsofun]

I hereby agree to the terms of the CLA available at: https://docs.databend.com/dev/policies/cla/

Summary

• Closes bug: correlated EXISTS with cross-type join keys (INT = DECIMAL) wrongly matches NULL probe keys — and returns an unstable number of rows #20364
• Add an exact regression for correlated BIGINT = DECIMAL(20, 0) join-key coercion
• Verify that three NULL probe rows do not falsely match the build-side decimal zero
• Cover both EXISTS and NOT EXISTS under the experimental and legacy hash joins

Motivation

PR #20354 fixed correlated hash joins that dropped nullable probe-key validity and consequently treated NULL fixed-width keys as their physical default values. Its regression tests cover same-type DECIMAL, BOOLEAN, and TIMESTAMP keys, but #20364 specifically reported a cross-type numeric comparison where casting a nullable BIGINT probe key to DECIMAL exposed the same defect.

This test-only follow-up preserves the issue's exact cardinality-sensitive input: three NULL rows plus two non-matching BIGINT values on the probe side and a single DECIMAL zero on the build side. EXISTS must return zero rows, while NOT EXISTS must retain all five probe rows. The assertions pass with #20354 and protect both hash join implementations from regressing this coercion path.

Tests

• Unit Test
• Logic Test
• Benchmark Test
• No Test - Explain why

Validation performed:

• cargo build --bin databend-query --bin databend-meta --bin databend-sqllogictests
• databend-sqllogictests --run 'tests/sqllogictests/suites/query/subquery.test' (273 MySQL and 273 HTTP tests passed)
• git diff --check

Type of change

• Bug Fix (non-breaking change which fixes an issue)
• New Feature (non-breaking change which adds functionality)
• Breaking Change (fix or feature that could cause existing functionality not to work as expected)
• Documentation Update
• Refactoring
• Performance Improvement
• Other (please describe): Test-only regression coverage for fix(query): preserve null validity in correlated joins #20354 and bug: correlated EXISTS with cross-type join keys (INT = DECIMAL) wrongly matches NULL probe keys — and returns an unstable number of rows #20364

AI assistance

• AI usage: An AI coding agent reproduced the cross-type case, drafted the focused logic-test coverage, and ran the listed validation commands
• Responsible human: @youngsofun
• The responsible human has read every line of this diff and can explain each change

---

This change is 

[sundy-li]

Verified: applied the diff on main, ran databend-sqllogictests against subquery.test with both mysql and http handlers (273/273 passed each). Test-only change, additive, follows existing file conventions. LGTM.