Maximilian Roos

> @max-sixty, is this a valid query? > > ```elm > derive a = 5 > ``` I could see two options: - We start with a null tablespace /...

> My concern is that derive only adds _columns_ not _rows_. Yes good point. Agree with option 2! Meta point — I think it's good that we're starting to hit...

> @max-sixty now, the semantic branch should be quite stable in the sense most of my major changes have been applied. I still have plans to change-up materializer (which will...

> I was doing that a lot, but `cargo build` and rust-analyzer still had long compilation times. For some reason compiling duckdb too way too long and sometimes used >5GB...

@aljazerzen how do you want to handle merging `main` in here? I'm guessing merge difficulty is super-linear in size — doing lots of small merges is easier than one big...

I did the full merge and ignored failing tests. Let's try and keep the changes here as scoped as possible — I appreciate that the place we're working is an...

> If we don't merge it will be easier to rebase/squash in the end when we are done with semantic. Sorry @aljazerzen — I just saw this and the message...

Unfortunately I also may have made this mistake on not seeing your message in time: > Make sure not to accept any failing insta snapshots, because they (mostly) fail for...

Thanks for starting this! I would split the feature into: - Static column names, known at compile time — we could write this in PRQL without too much trouble -...

> Am I correct in assuming that pivot operation requires dynamic columns, while melt produces only a static columns (independent of table contents)? I think: - With melt, the query...