Zhang Chao

Thanks for your suggestion, i'm thinking about that too, i will continue working on this.

Sorry for the late update! i spent some more time on other PR, i encoded the positions with group-varint when `storeOffsets` is false and there are no payloads. with the...

> i'll test it with #12841 optimized code tomorrow. emmm... there's still no significant performance improvement, possibly the write path `writeGroupVInts` is a bit slower than `writeVInt`, because group-varint needs...

Hi Adrien, for simplicity, I changed the logic to use group-varint for when positions only (no offsets and no payloads). In addition, since `writeGroupVInt` is slower than `writeVInt`, we don't...

Yeah, it looks like we can optimize `writeGroupVInt` in the same way as we did for the read logic. I'd love to try and run some benchmarks.

@jpountz Thanks very much for the guidance! Your explanation makes sense to me, I'll try this idea :)

Hi Adrien, i tried the idea, this will got a ~10% speedup on JHM output, which is slightly less because it needs an extra loop to decode the delta. The...

It looks like you are using a lower version of lucene to read the index generated by a higher version. Are you sure the version with this exception is lucene...

Yeah, I’ve been thinking about this. Elasticsearch now supports a time_series index mode with DELTA + FOR encoding on doc values. In time series or logging scenarios, storage cost usually...

@rmuir You are right, it needs to be sorted on the timestamp field. In addition to enabling delta-compression on the timestamp field, index sorting brings another benefit: when sorting by...