Use relationship on dplyr >=1.1.0.9000

[DavisVaughan]

We are preparing to release dplyr 1.1.1 and this package popped up in the revdep checks.

We realized that we didn't get the multiple argument quite right the first time around 😞 . It was too aggressive since it warned on both one-to-many and many-to-many joins. In 1.1.1 we've made two improvements:

• multiple now defaults to "all". The options of NULL, "error", and "warning" are deprecated.
• relationship is a new argument to add known constraints onto the join procedure, such as "many-to-one", which replaces multiple = "error".

The default of relationship checks to see if there is a many-to-many relationship between the keys of x and y and will warn if one is present. This should be much rarer than what we checked for before, and targets the most dangerous case that we were trying to warn the user about.

You can read all about relationship here tidyverse/dplyr#6753, along with the issues linked there.

Unfortunately it does affect some tests here. I think the easiest thing to do is to just branch off the dplyr version that is installed and set relationship = "many-to-many" or multiple = "all" accordingly.

We plan to submit dplyr 1.1.1 in 2-3 weeks.

This should be compatible with both dev and CRAN dplyr. It would help us out if you could go ahead and send a patch version of your package to CRAN ahead of time! Thanks!