Maximilian Roos

> What about just having `exclude`/`except` transform? Yes, no strong view against it specifically. Keeping the cardinality of keywords low means the language doesn't become cluttered. I think we applied...

> ``` > select [*] except:[foo, bar] > ``` > > ? Yes +0.5!

Or: ```elm select (* -[foo, bar]) ``` ?

> select [^ foo, bar] One downside with this is `^` seems to apply to the first item, but actually applies to all the items. I think this would be...

And to be explicit we can have > ```elm > select ![foo,bar] > ``` but not `select -[foo,bar]` because `!` is not a binary operator, so it can't be confused...

This was discussed at https://news.ycombinator.com/item?id=39717268. We should ensure that this is at least possible with an s-string.

That's a great list — it does look encouraging. Though the projects all look like they also support SQL too for the moment. I'm mood-affiliated to support this, though until...

Good spot. We can add a replacements like https://github.com/prql/prql/blob/main/prql-js/Cargo.toml#L36-L53 to the `prql-java` (although no impact until we do a pass of the java release...)

It seems like resolving is fixed — but in the translation, to what extent do we need to quote the keywords in the resulting SQL? i.e. ```diff SELECT call_id, timestamp,...

Yes good point. But then at the translation — when there's an ident (i.e. there are no functions any longer) — then shouldn't be quote keywords, in the same way...