Repository navigation
Conversation
For this purpose, convert all input elements using vectorize rather than nditer, to avoid coercion to arrays and thus allow Column.__getitem__ to do conversion from bytes to string.
|
Hi there @mhvk 👋 - thanks for the pull request! I'm just a friendly 🤖 that checks for issues related to the changelog and making sure that this pull request is milestoned and labeled correctly. This is mainly intended for the maintainers, so if you are not a maintainer you can ignore this, and a maintainer will let you know if any action is required on your part 😃. Everything looks good from my point of view! 👍 If there are any issues with this message, please report them here. |
|
@taldcroft - on the speed: I checked with a list of 10000 strings, and this PR is about 25% slower than master. |
Hmm, that's not quite negligible. It's a bit ugly, but this might be worth two distinct code paths: one which is "faster" and works for unicode and one which does the decoding for bytes input. Again, one knows up front the type of the data. Or maybe there is a clean way to not lose speed in the default case of unicode inputs? |
|
@taldcroft - I think I now agree with you that my original, more liberal attempt, was better. So, let me close this one... |
For this purpose, convert all input elements using vectorize rather than nditer, to avoid coercion to arrays and thus allow Column.getitem to do conversion from bytes to string.
This is an alternative to #6823, probably better as less intrusive and more targeted.
This is helpful for #6821.