Recipes documentation unclear on how to pass data for CV

[jxu]

In ?recipes, the data passed to recipes() in the example is the training split. But it is unclear what should be passed if we split data using vfold_cv. Is it the entire dataset? The docs seem to suggest the actual data doesn't matter for the recipe, just the column names.

https://www.tidymodels.org/start/resampling/ does not show how to use a recipe, just formula.

[EmilHvitfeldt]

We typically don't use a vfold_cv fold object directly with a recipe() object. packages like tune knows how to handle both recipes and vfold_cv objects.

The dataset passed to recipe() does a couple of things. It is used to denote the column names, and their types. This information is used to detect whether some variables are missing, or if they have the wrong types.

When you go to prep() a recipe it will use that data by default, if the training argument of prep() has not been set.

[jxu]

So only the column names and types of the data passed in matter? It's not clear in the documentation.

[github-actions]

This issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with a reprex https://reprex.tidyverse.org) and link to this issue.