Phillip Cloud
Phillip Cloud
@vspinu You can use `ibis.memtable(df)` and Ibis will automatically create a temp view on your behalf. No need to call into the PySpark API for that particular use case!
Yep, I misread the comment. Returning a PySpark DataFrame directly is probably out of scope, but you should at least be able to turn one into an Ibis table easily.
@leofang It looks like there's already a `cuda-cccl` package (https://github.com/conda-forge/cuda-cccl-feedstock). What needs to be done for this issue to be resolved?
This is done.
Can you try reading a file using vanilla `pyarrow`?
This requires a lot more work. Because of the way Ibis models window functions, a backend that wants to support them must support compiling `ops.WindowFunction` first, and **then** implement support...
There's no guarantee that things would have a particular name, because we might move expressions around or change them in some way that would break anything a user specified here....
Removing `get_name` from the public API seems like the best compromise here, given that it's an implementation detail.
Seems to be failing more than half of CI checks.
Given that grouping by `UNNEST` works in postgres, I'd much rather get this implemented upstream in DuckDB than hack around it here.