`RuntimeError` should be warning

[LilianBoulard]

When passing an already "clean" dataset to the SuperVectorizer (which is common in generic pipelines), it raises a RuntimeError as No transformers could be generated!. The doc mentions

Raises
------
RuntimeError
    If no transformers could be constructed,
    usually because transformers passed do not match any column.
    To fix the issue, try passing the least amount of None as encoders.

First, I think the message could be reworded, as it's not very clean what "no transformers being generated" means. Secondly, I think it should be a warning (at best), not an error.

[GaelVaroquaux]

Yes, we shouldn't error on clean datasets, just pass them along. I'm not sure that a warning is warranted.

On Aug 5, 2022, 11:26, at 11:26, Lilian @.***> wrote:

When passing an already "clean" dataset to the SuperVectorizer (which is common in generic pipelines), it raises a RuntimeError as No transformers could be generated!. The doc mentions

Raises
------
RuntimeError
   If no transformers could be constructed,
   usually because transformers passed do not match any column.
   To fix the issue, try passing the least amount of None as encoders.

First, I think the message could be reworded, as it's not very clean what "no transformers being generated" means. Secondly, I think it should be a warning (at best), not an error.

-- Reply to this email directly or view it on GitHub: https://github.com/dirty-cat/dirty_cat/issues/296 You are receiving this because you are subscribed to this thread.

Message ID: @.***>