Phillip Cloud
Phillip Cloud
@Riezebos Any interest in submitting a PR? Adding this to the `memtable` docstring somewhere would be the place to start!
Looking into this, a large source of the Ibis overhead here is the 450 relations we are constructing, each of which must go through binding and dereferencing neither of which...
For the curious, here's the DuckDB plan where you can see the `UNNEST`: ``` D create table penguins as from read_csv('/data/penguins.csv'); D explain summarize (select year from penguins); ┌─────────────────────────────┐ │┌───────────────────────────┐│...
A follow up to #9684 might actually make a proper `Summarize` operation in Ibis that at least in the DuckDB case would have an extremely small constant overhead to construct...
Even before that, we could probably move this entire thing to compilation time which would make this much cheaper for both the union case and the array case ... I'll...
Seems reasonable!
@KeXiangWang Are you still interested in this?
@hottwaj It looks like you may need to remove the xfail markers on the tests that are now `XPASS`ing.
Which is pretty nice, it means there were already a couple of tests that hit this code path :)
@hottwaj Who is @jc-5s? Ideally the people that submit PRs are the same people who are making commits.