Patrick Hoefler
Patrick Hoefler
Hi @thomas-fred thanks for providing the reproducer. We have a fix here: https://github.com/dask/dask-expr/pull/1000
Sorry about that, https://docs.dask.org/en/stable/dataframe-api.html we switched the link
@kritikakshirsagar03 Could you add a bit more context what you want to tell us with this?
This is a little odd and looks to be related to how we tokenise keys.
thx for sticking with this! Feel free to pick other issues if you are interested
Thanks for your report. This basically follows the broadcasting rules of Numpy, i.e. we are broadasting to the second dimension: ``` arr = np.arange(0, 20).reshape(5, 4) arr + np.arange(0, 4)...
Closing this as expected behavior, it's also documented in the docstrign a little further down that we are aligning blocks along every dimension
Having a Dask DataFrame that has mismatching columns can cause all kinds of issues. I would definitely not recommend doing this. I don't think that read-json has this built in,...
https://github.com/dask/dask/issues/11190#issuecomment-2180835903