Andrey Vasnetsov
Andrey Vasnetsov
> Qdrant searches each shard (or segment) independently and in parallel. With auto sharding, it queries all 6 shards in parallel and then aggregates the results. With custom sharding, we...
> Would it be possible to schedule a meeting to discuss this further? Yeah, sure, please contact our sales or share your contact information so our sales would contact you....
> if I try to search only with sim search I get much better results with scores 0.8+ maybe there are no results matching the filter in your dataset? You...
fusion overwrites scores
is fusion results are not part of both pre-fetches? Please ignore absollute value of the score
Hey @woaiqjj, thanks for reproring and reproduction instructions. Qdrant indeed accepts an extended version of RFC3339, which is described here https://qdrant.tech/documentation/concepts/payload/?q=datetim#datetime This extension is done for convenience purposes and should...
hey @vincent-pb Thanks for the proposal! Could you please describe in a little more details how those additional fields would help you? Is that purely for debugging purposes?
I don't think providing reliable ETA is even possible, assuming we don't know in advance the size of each segment which needed to optimize. We do currently have a `"log"`...
Also, I don't think we can provide global progress, as indexation might start before all points are uploaded, so we are risking being very noizy and unreliable regarding global indexation...
Let's keep this issue open, so we will notify you about the progress on improving