Rearranging output of `inspect_cat()` removes labels for `show_plot()` about inspectdf

[alastairrushworth]

Thanks again for the original report. I've now pushed a change that should allow sorting and filtering of the outputs of inspect_cat() to be respected by the downstream show_plot() call.

[alastairrushworth]

Hi @RoelVerbelen. I think the issue of using arrange() causing the downstream plot to have distorted label was fixed by a previous commit. Please let me know if it hasn't.

You also point out that if jsd is NA that the column comparison is dropped from the graphic. I'm curious about the cases that might have caused this to happen - did you see a specific instance of this? One case where this happens is where the column has different types in the two data frames, in which case I think dropping from the graphic could make sense. But that may not be true in general...

Once again, thanks for the feedback and reports 😊

[RoelVerbelen]

Thanks again @alastairrushworth. In latest commit version, the labels are back but the arranging is completely ignored. Plotting is always done with the variable names sorted alphabetically. What would be the recommended approach to reorder the variables on the y-axis based on their corresponding jsd value?

I don't remember unfortunately how I ended up getting missing values for jsd before. I tried creating examples for the starwars data where one version of the data has 0 observations for some factor levels, thinking it would cause a division by 0 or log(0) in the jsd formula, but these scenarios seem to set it to 1. Also the trivial example where a factor is always missing resulted in a jsd of 1. I'll create a new issue if I run into missing values for jsd again.