OpenSearch version 2.5
I would like to build and index based on binary data, and use hamming distance for approximate search.
From what I saw there is only support for exact search (script score) using hamming distance (hammingbit), but it takes too much time…
Is there any other way to achieve KNN approximate search?
I am familiar with this tutorial, but from here is seems that it is indeed not supported.
Wanted to validate if this is indeed the situation, and if there is any plan to support it?
Oh, I’m sorry, I misread your question. I don’t think there’s support for it and I’m not sure how this can be added to be more efficient than a script score (which sounds expensive indeed).
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.