branch-4.1: [feature](variant) Support ColumnVariantV2 segment reads and writes - #66607
branch-4.1: [feature](variant) Support ColumnVariantV2 segment reads and writes#66607eldenmoon wants to merge 2 commits into
Conversation
|
Thank you for your contribution to Apache Doris. Please clearly describe your PR:
|
|
run buildall |
BE Regression && UT Coverage ReportIncrement line coverage Increment coverage report
|
FE Regression Coverage ReportIncrement line coverage |
|
run buildall |
BE UT Coverage ReportIncrement line coverage Increment coverage report
|
|
run buildall |
|
run beut |
BE UT Coverage ReportIncrement line coverage Increment coverage report
|
BE Regression && UT Coverage ReportIncrement line coverage Increment coverage report
|
FE Regression Coverage ReportIncrement line coverage |
### What problem does this PR solve? Issue Number: None Related PR: apache#66204 Problem Summary: Backport PR apache#66204 to branch-4.1 while preserving Variant V1 behavior when Config.enable_variant_v2 is disabled. Native Doris VARIANT now follows the FE config exclusively, Paimon uses the same global policy, the removed session variable no longer participates in V1/V2 routing, and computeV2 remains only as the physical-format marker required by external Iceberg and Paimon paths. Paimon scalar regression expectations are aligned with the existing Doris SQL serialization contract. ### Release note Support ColumnVariantV2 segment reads and writes on branch-4.1. Native Doris and Paimon V1/V2 routing follows Config.enable_variant_v2. ### Check List (For Author) - Test: Regression test / Unit Test - BE Variant unit tests: 714 total, 709 passed and 5 preset skips - PaimonVariantWriteAnalyzerTest: 9 passed - PaimonScanNodeTest and VariableMgrTest: 52 passed - DataTypeVariantV2SerdeOutputTest: 7 passed - variant_p0 with Config=false and Config=true: 163 passed in each mode; one OSS credential-dependent case was excluded due InvalidAccessKeyId - build-support/clang-format.sh - Behavior changed: Yes. Variant V2 reads and writes are enabled when configured; legacy V1 behavior remains unchanged when disabled; enable_variant_v2 is no longer a session variable. - Does this need documentation: No
be6ea02 to
f5286b9
Compare
|
run buildall |
1 similar comment
|
run buildall |
### What problem does this PR solve? Issue Number: None Related PR: apache#66204 Problem Summary: The mixed-row-group Parquet test rendered extracted Variant V2 values through the SQL to_string path. Top-level SQL strings intentionally follow the legacy V1 display contract and omit JSON quotes, so the test no longer verified its expected JSON representation after the backport. Serialize the extracted values through the explicit JSON API so the assertion remains type-sensitive without changing SQL behavior. ### Release note None ### Check List (For Author) - Test: Unit Test - NewParquetReaderTest.SwitchesVariantLeafProjectionPerRowGroup - DataTypeVariantV2SerdeOutputTest.* - Behavior changed: No - Does this need documentation: No
|
run buildall |
FE UT Coverage ReportIncrement line coverage |
FE Regression Coverage ReportIncrement line coverage |
cherry-pick #66204