Feature request: Suggestion to rename tbl_strata_nested_stack()/tbl_strata() to clarify long vs. wide stratification output

[martincadek]

Currently tbl_strata() (with default .combine_with = "tbl_merge") produces strata side-by-side as columns, while tbl_strata_nested_stack() stacks strata as rows. The current names don't immediately convey this "wide vs. long" distinction to users unfamiliar with the internals.

Suggestion: Consider renaming (with a standard deprecation cycle, e.g. lifecycle::deprecate_soft()) to:

tbl_strata() → tbl_strata_wide()
tbl_strata_nested_stack() → tbl_strata_long()
This would mirror the well-known wide/long terminology from tidyr::pivot_wider()/pivot_longer(), which many R users already have as a mental model for row- vs. column-oriented reshaping.

Open question: Since tbl_strata() supports .combine_with for both merge and stack behaviour, would it make more sense to keep tbl_strata() as the general-purpose function and only introduce tbl_strata_long() as an alias/wrapper for the stacking case, rather than renaming the base function?

[ddsjoberg]

Thanks for the post @martincadek

The tbl_strata() function has been around for a very long time, I think it would cause quite a few issues to rename it as this point.

Since tbl_strata() supports .combine_with for both merge and stack behaviour, would it make more sense to keep tbl_strata() as the general-purpose function and only introduce tbl_strata_long() as an alias/wrapper for the stacking case, rather than renaming the base function?

This could be a way forward. But we also have tbl_strata2(), so we would end up with 4 new 'helpers' that save the user for changing one argument. I am not sure if that is the best use of the NAMESPACE. Perhaps you could consider adding these helpers to crane first. If they end up being quite useful, we could consider exporting them here instead