Matthew Rocklin

Thank you for the error report @ResidentMario . My apologies in the delayed response. The folks who maintain this repository have been busy lately. Do you have any interest in...

Thanks for raising this issue @mt-jones ! In general contributions would be welcome. I think the only thing we would want to be careful about would be explicit dependencies on...

Also, just to be clear, I'm looking forward to including the GPU support to this project. I think that it will be a great contribution. Please don't interpret my pickiness...

> The proper concat function needs to be called Dask dataframe handles this with the `dask.dataframe.methods.concat` function, which will call the appropriate function based on the inputs provided. > that...

> Are you suggesting that Dask can replace the logical concat here? Yes. I recommend taking a look at `dask.dataframe.methods.concat`, which has a dispatched solution. > Namely, XGBoost cannot be...

Imports only happen once per Python process ```python In [1]: %time import cudf CPU times: user 804 ms, sys: 1.45 s, total: 2.25 s Wall time: 4.51 s In [2]:...

In a side channel conversation @mt-jones asks > If you could provide me a sample of the method you referred to in order to abstract the `concat` method, that would...

I would nuke the concat function defined in this repository and replace it with `dask.dataframe.methods.concat`. If there are cases not handled by `dask.dataframe.methods.concat` then I would add them to that...

Currently there is no easy workaround. It would be good to provide one though. I would think that this work would happen upstream in dask.dataframe , but I don't have...

I think that all of the boosters are effectively the same, so we only need tor return one of them. On Fri, May 31, 2019 at 7:08 AM Tom Augspurger...