Guo Feng

LUCENE-10409: Improve BKDWriter's DocIdsWriter to better encode decreasing sequences of doc IDs

2

https://issues.apache.org/jira/browse/LUCENE-10409 #11445

LUCENE-10366: Override #readVInt and #readVLong for ByteBufferDataInput to avoid the abstraction confusion of #readByte.

32

https://issues.apache.org/jira/browse/LUCENE-10366 #11402 Today, we do not rewrite `#readVInt` and `#readVLong` for `ByteBufferIndexInput`. By default, the logic will call `#readByte` several times, and we need to check whether `ByteBuffer` is valid...

LUCENE-10376: Roll up the loop in vint/vlong in DataInput

11

https://issues.apache.org/jira/browse/LUCENE-10376 #11412

LUCENE-10334: Introduce a BlockReader based on ForUtil and use it for NumericDocValues

4

see https://issues.apache.org/jira/browse/LUCENE-10334 #11370 Changes: 1. Make ForUtil fit for 64bit. 2. Introduce the new `BlockWriter` and `BlockReader`. 3. Replaces `DirectReader` with `BlockReader` and `DirectReader` with `BlockWriter`.

Lucene-10334: Detected warm up of directreader

2

see: https://issues.apache.org/jira/browse/LUCENE-10334 #11370

LUCENE-10333: Speed up BinaryDocValues with a batch reading on LongValues

10

See https://issues.apache.org/jira/browse/LUCENE-10333 #11369 PS: I'm pushing these codes to quickly see if this approach makes sense to you, I'll add some more tests if this can pass the the first...

TermInSetQuery will avoid caching queries using too much memory, but there is no such limit in PointInSetQuery and PointInSetQueryWithScore, which may have similar problems. I added the logic to these...

Generalize LSBRadixSorter and use it in SortingPostingsEnum

9

**Description** In https://github.com/apache/lucene/pull/12114, we had great numbers for LSB radix sorter when sorting random docs in `SortingDocsEnum` . But we can not take advantage of the LSB radix sorter in...

Move index of numbers from BKD into DocValuesSkipper?

2

### Description Immature idea, for discussion purpose :) Today our range query on numbers have two approaches: BKD or DocValue Skipper. The drawback of BKD: * Allocating too much memory...

Compression cache of numeric docvalues

7

### Description When benchmarking recently with some OLAP engines (no indexes, no stored fields, only column data), the results showed that they only occupy 50-70% of the storage of `NumericDocvalues`,...