Fork me on GitHub
Slack Archive

This page is not created by, affiliated with, or supported by Slack Technologies, Inc.

2022-06-03

Channels

#xtdb
<
2022-06-03
>
tatut06:06:30

I took a cursory glance at this kv store open sourced by vmware https://deluxe-bunny-5b210c.netlify.app/posts/introducing-splinterdb/ does that look like it could be used as an index store?

šŸ‘€ 1
Hukka06:06:34

That domain nameā€¦

tatut06:06:39

http://splinterdb.org is the main one but redirects to those

Hukka07:06:15

> SplinterDB surpassed RocksDB by up to seven times in insertions, four times in reads, in device bandwidth utilization, and on non-random workloads performance sure sounds interesting

refset20:06:34

> does that look like it could be used as an index store? Late to the party...but it sounds like it, yup šŸ™‚

Dan14:06:16

Hi everyone, so XTDB breaks down vectors into individual components which breaks our desired functionality for limit and offset We use collection binding to pass a vector to our query and we get multiple results for a single xt/id. We then have to transform this result into one document per vector If we dont use collection binding we only get back one of the components that XTDB stored from the document with a vector How can we query a document that contains a field with a vector and only return one document per match so that we can use limit and offset?

Martynas Maciulevičius15:06:34

I'm not sure if I understand you question but I think you want to do this to filter by the non-expanded list [1 2 3]: 

{:top-level-attr [[1 2 3]]}
Then you would probably want to sort the result and do limit+offset as you do normally. And in this [1 2 3] won't be indexed and expanded: 
{:top-level-attr {:nested-unexpanded-list [1 2 3]}

šŸ™ 1
Dan15:06:47

Thank you so much for your quick response, i was able to remove collection binding and using the first example seems to be what i want

āž• 1