The future of top_n()?

[ofchurches]

Thank you so much for this package. I have used it so much over the last two years and will keep using and recommending it.

I have a frequent workflow that selects nodes with the highest value on some measure:

library(tidygraph)
iris_clust <- hclust(dist(iris[1:4]))

as_tbl_graph(iris_clust) %>%
  top_n(n = 10, wt = height)

I note that top_n() is reexported from dplyr. However, since dplyr release 1.0.0, top_n() has been superseded in favour of slice_min()/slice_max(). However, there appears to be no applicable method for slice_min()/slice_max() applied to tbl_graph objects created by tidygraph.

I see some changes due to dplyr 1.0.0 in tidygraph already (#119) and I can see good reasons either to change tidygraph verbs for consistency with dplyr or to keep tidygraph heading on its own path. I also appreciate that top_n() and slice_min()/slice_max() have interesting impacts on tbl_graph objects because they remove the edges terminating at the removed nodes and change the indexes of the to and from columns in the edge data. So this is a non-trivial job.

So, could I ask what your thoughts are about the future of top_n() in tidygraph?

It would really help me plan my code for the future to know how this is going to evolve.

Again, many thanks for your work. I really love this package.