[Caideyipi]

Description

WAL deletion replay currently writes mods that can also hide inserts replayed later into the same recovered file. Restrict recovered mods to each flushed device and its on-disk end time while preserving the original FIELD and time ranges, tree wildcard semantics, table isolation, and NULL/empty TAG identities.

Correct table last-cache handling for sparse FIELD values, last_by(TAG/ATTRIBUTE, Y), NULL identities, deleted devices, time predicates, TTL, and grouped/partial aggregation. Query writeback now uses a dedicated entry that updates the device row time only from an explicit TIME result; existing write-side cache updates retain their behavior. Apply the same contract to direct and cross-Region query writeback, and avoid modifying shared cached pairs when filtering query results.

Add focused recovery and last-cache regressions. Unrelated deletion/compaction ITs already in #19051 are excluded. The identical DeletionTestUtils helper is retained because the recovery regressions require it for this PR to compile independently against master.

Validation

• Independent workspace re-review completed with no findings. The retained production code, unit tests, and recovery/last-cache test logic match the reviewed snapshot; publication only removes the unrelated duplicate test blocks and unused imports.
• Retained execution evidence: Windows/JDK 17 with TsFile 8f665d113dbb5257b901c528c31c8bc24da2a018, without version mapping; clean DataNode/IT package, 12 cache unit tests, and 9 existing last-cache ITs plus 2 isolated verification-only probes passed. The latest last-cache cluster run used 1 ConfigNode and 3 DataNodes. Recovery ITs also have earlier passing evidence.
• The 3 added TsFilePlanRedoerTest methods have passing remote evidence and have not been run locally. Two pre-existing ignored tests remain unexecuted. The full IT suite was not run; builds used -Denforcer.skip=true.
• Publication checks: Spotless formatting, IT discovery/shard selection, scoped diff checks, and exact retained-source comparison. Build/test results above were reused from the reviewed snapshot at 28c5028bb462403315e46319b57e9beb96ffaa69; compilation and executable tests were not repeated after refreshing master and removing duplicate ITs.

This PR has:

• been self-reviewed.
• added unit tests or modified existing tests to cover the fixes.
• added integration tests.
• been tested in a test IoTDB cluster as described above.

[sonarqubecloud]

Quality Gate failed

Failed conditions
C Reliability Rating on New Code (required ≥ A)

See analysis details on SonarQube Cloud

Catch issues before they fail your Quality Gate with our IDE extension SonarQube for IDE

[codecov]

Codecov Report

❌ Patch coverage is 39.53488% with 52 lines in your changes missing coverage. Please review.
✅ Project coverage is 48.11%. Comparing base (1b2158f) to head (93dec7a).

Files with missing lines	Patch %	Lines
...urce/relational/LastQueryAggTableScanOperator.java	0.00%	35 Missing ⚠️
...e/plan/planner/DataNodeTableOperatorGenerator.java	0.00%	7 Missing ⚠️
...metadata/fetcher/cache/TableDeviceSchemaCache.java	0.00%	4 Missing ⚠️
.../metadata/fetcher/cache/TableDeviceCacheEntry.java	0.00%	3 Missing ⚠️
.../dataregion/wal/recover/file/TsFilePlanRedoer.java	93.75%	2 Missing ⚠️
...ngine/execution/fragment/DataNodeQueryContext.java	0.00%	1 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff              @@
##             master   #19081      +/-   ##
============================================
- Coverage     48.12%   48.11%   -0.01%     
  Complexity      625      625              
============================================
  Files          5572     5572              
  Lines        411488   411537      +49     
  Branches      54166    54185      +19     
============================================
+ Hits         198010   198017       +7     
- Misses       213478   213520      +42     

☔ View full report in Codecov by Harness.
📢 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.