Corneliu Cofaru

> @MN-company There's no news yet. If I understand the context correctly, @ddvk is afk (vacation?). As he is sole maintainer, **version 2.9 is not expected until @ddvk is back**....

Two issues: - there seems to be a bug for the inner join for `DNDSparse`; `NDSparse` join works and can be tested with `join(Ascalar, Bscalar)` - for a join, ideally...

I suppose this would remove the possibility of in-place modifications on the container as the iterator would yield values rather than references. So, just to exemplify, the curent behaviour allows...

I have to say I like the current behaviour better, for its consistency (i.e. when iterating, yield subsets, call `getobs` to make a copy).

This type of iteration in streams seems powerful but the concept of observation becomes somewhat ambiguous, as well as dimensionality. Is there a specific usecase you are refering to ?

I based the ambiguity statement on the assumption that the data is not in any collection-like form but a stream/buffer (where calling `nobs` may or may not make sense i.e....

Indeed, I see the peeve. The concept of randomness does require the knowledge of some temporary `N` (available data at some moment) or something on that line in other dimension....

I can attest that [nmslib](https://github.com/nmslib/nmslib) is extremely fast - takes on the order of a few milliseconds to perform a search in millions of samples for a dimensionality >100. That's...

Thanks @Datseris for the links awesome :) Reffering to #69, the simplest (and maybe fastest) approach would be to implement a minimalistic ann directly in Julia ... I will look...

@JonasIsensee great, I can help with testing and further optimization if needed.