fdb-record-layer

fdb-record-layer copied to clipboard

Within the Cacades planner, there are methods for converting certain `EXPLODE` expressions into `InSource`s for IN-join and IN-union planning. However, this logic is restricted to only `LiteralValue`s (which get turned...

alecgrieser

For CJK, a prefix search like AB* would not return ABC since Lucene does not run AB* through an analyzer. For this, we would need to split into (A and...

qdnguyen0

This assertion appears to be failing intermittently for the `OnlineIndexerBuildRankIndexIndexTest::addWhileBuildingParallelRank` test: https://github.com/FoundationDB/fdb-record-layer/blob/2f127f62e77a71d0450c948100735dc457b74d23/fdb-record-layer-core/src/test/java/com/apple/foundationdb/record/provider/foundationdb/OnlineIndexerBuildIndexTest.java#L301 For example, this test failure on commit hash 396a0ce52f786460e75d24d420b11657b0bf58ca: ``` OnlineIndexerBuildRankIndexTest$Unsafe > addWhileBuildingParallelRank(long) > com.apple.foundationdb.record.provider.foundationdb.OnlineIndexerBuildRankIndexTest$Unsafe.addWhileBuildingParallelRank(long)[2] FAILED org.opentest4j.AssertionFailedError: expected:...

alecgrieser

Resolves #2801: Covering optimization with invertible index fields

5

This is complete for the heuristic planner. For Cascades, it adds the necessary partial record support, but more work in `tryFetchCoveringIndexScan` is needed, depending on how precisely function index keys...

MMcM

Add more complete direction control to LogicalSortExpression.

10

Extends #2748 and #2771.

MMcM

Add a test to check behavior of ordered union with mixed direction sorting

1

MMcM

An `InvertibleFunctionKeyExpression` should be in `AvailableFields` in its original (inverted) form and inversion should be supported by `IndexKeyValueToPartialRecordTest`.

MMcM

[SQL] Support Tuple IN LIST<Tuple>

3

pengpeng-lu

Fixes #2798: TextCollatorRegistryICU.getTextCollator should freeze the Collator

2

MMcM

Change representation of aggregate match candidates

7

This proposes a new representation of aggregate plans that gets rid of the first grouping Value in the underlying select expression, and consumes @normen662's fix for an issue with `Value`...

hatyo