Closed jona-sassenhagen closed 2 years ago
@janmotl any chance you can review this?
@jona-sassenhagen Can you confirm that the new call also works in the oldest version of Pandas that category_encoders claim to support? If yes, I do not see a reason why the PR should not be merged.
However, I am not active in the project anymore and the merge must be done by some of the current maintainers. @zachmayer ?
Yes, the new call should work, there was no change to inplace
since .21: https://pandas.pydata.org/pandas-docs/version/0.21/generated/pandas.DataFrame.drop.html
Who are the current maintainers? I don't want to randomly ping people. Zach isn't a maintainer I'm pretty sure :)
Thanks!
@jona-sassenhagen I don't know.
Hi Jona,
thanks for your contribution. I'm one of the maintainers. I'm sorry it always takes some time until one of us reviews stuff.
The PR looks good to me and I'll merge it when the pipeline succeeds.
It seems like our github runner instance does not support Python 3.5 anymore. So the problem seems to be on our side, not on yours. I thinks we should deprecate python 3.5 anyways since it ran out of the 5 year support period. Please bear with me with this issue.
Sure no worries. Thanks a lot.
@jona-sassenhagen I've updated the pipeline. Can you please pull the latest master so that I can run the checks again?
@PaulWestenthanner like this?
🙏
When I use
drop_invariant=True
, I get aFutureWarning
becausedf.drop
is switching to named kwargs only. This fixes it.It should also be substantially more performant as the DF is only modified once.