Hi, looking at the issue, there was also first point regarding the default value of percent10. Could you also change the documentation and perhaps also the default value in _fetch_brute_kddcup99 helper function ? Thanks !

Sure, I'll update the pull request to reflect those changes as well.

I think it makes sense to leave the default value of percent10 as False (so calling the function without specifying the parameter downloads the complete data).

Let me know if you think this should be the other way and I can update the PR. Otherwise, I think it is good to go.

Hi, thanks for updating this. I think there was a discussion at #5946 here for changing it to True. But I suppose one of the core devs can give a better idea.

Good point, if one of the core devs can chime in about which default value it should have, I can update it.

LGTM

Actually reading #5946 (comment) it seems like the idea is to have percent10=True by default. I tried and the data is indeed quite big. On my machine it took all the RAM I had left (~14GB) and my system started swapping.

I reckon the documentation should be changed rather than the default value. As @maniteja123 said if you could change _fetch_brute_kddcup99 to have a consistent default and documentation for percent10, that would be great.

Updated!

Thanks a lot, merging!