hodlbod

It does cover the GROUP BY case popularized by sql, yes. Just like limit = limit, since/until = order by, offset, kind/#e/#p/author = where. Grouping is useful.

@fiatjaf I see what you mean though, this opens the door to more and more complexity, while second-layer protocols or centralized services can perform this function successfully. I'm just not...

Alright, since groups seem to be unpopular I've removed them, please take another look.

That's a super good point, I hadn't thought of that. I do think COUNT could still be useful as an approximation (since it is anyway), you could take the max...

Aaand I've changed my mind, I've decided that COUNT is not useful. If we add [relay extensions](https://github.com/nostr-protocol/nips/pull/259) we could put something together that would be more complete.

> I'm adding COUNT to my metadata / contact indexing nodes This is exactly the place that COUNT makes sense. It's not good for direct use with multiple relays, but...

I've re-introduced `group_by` to fit @adamritter's implementation. If relays prefer not to implement it, they can raise a NOTICE or ignore the group_by key. I think we should go ahead...

Can you elaborate? Currently coracle only uses COUNT in one place, but it allows me to avoid downloading many megabytes of data to populate a single number. A couple use...

Closing in favor of #746 and #686

Ok, that all sounds fine, I'm still not sure I see the semantic difference between the two, but if other people find kind 2 useful I'm happy to leave it...