fdb-record-layer

fdb-record-layer copied to clipboard

In the Cascades planner we try to reuse references if possible. We should not ever reuse references that can be pruned later on.

normen662

On some databases, as it seems, dangling index entries are part of the normal flow. To support routine cleanup of these entries, we should add: 1. `OnlineIndexScrubber` API argument to...

jjezra

This proposes a new representation of aggregate plans that gets rid of the first grouping `Value` in the underlying `select` expression. Doing so would achieve two goals: - Avoid ambiguous...

hatyo

predicate push down

4

normen662

change modeling of `OrderingPart` to use a sort order instead of a boolean for reversedness

1

This is related to #1318. It would be useful to state that certain ordering parts are fixed, i.e. bound to one value vie an e.g. equality predicate. In those cases...

normen662

There appears to be some amount of flakiness in the `LuceneIndexTest` that can result in `transaction_too_old` errors. For example, this one was hit during the prbs for #2742: ``` LuceneIndexTest...

alecgrieser

It’s possible for the query to be MultiPhraseQuery in buildPhraseQueryWithPrefix when the index analyzer generates multiple tokens per position Example: WordDelimiterFilter with PRESERVE_ORIGINAL, Synonym Analyzer, NGRAM Analyzer

qdnguyen0

WIP - recursive CTE

2

hatyo

Some followup from PR #2954

2

- [Add MoreAsyncUtil.ALREADY_CANCELLED](https://github.com/FoundationDB/fdb-record-layer/commit/aaf4ebd4796ff08a6795b9fc87f749a2e1fc49cd) MapPipelineCursor and FlatMapPipelinedCursor had a version of this, so reducing duplication. - [Test concurrent closing with a non-singleton inner cursor](https://github.com/FoundationDB/fdb-record-layer/commit/d489748fcf00bfcf7352620f0a372a1688d89437)

ScottDugas

group by sets

2

normen662