Handle stat_bin2d(breaks=(x=NA,y=Something))

[tgs]

With this change, you can specify one set of breaks to stat_bin2d without setting the other. This is particularly useful when binning by date in one dimension, and using a transformed scale in the other dimension. This way, you don't have to figure out how to reconstruct the default breaks for the transformed scale. For example:

breaks <- list( x = sort(unique(movies$year)) - 0.5, y = NA )
g <- ggplot(movies) + aes(x = year, y = length)
g + stat_bin2d(breaks=breaks) + scale_y_log10()

[hadley]

I think using list(x = NULL, y = NULL) makes the code below a bit simpler. Then you can just test is.null(breaks$x)

[tgs]

Sure, sounds good.

[hadley]

Would you consider adding a couple of unit tests? I'd start with something like:

df <- data.frame(x = 1:3, y = 1:3)
p <- ggplot(df, aes(x, y)) + 
  stat_bin2d()
ggplot_build(p)$data[[1]]

[tgs]

Sure, I'll write some more specific unit tests. Actually, just got started, and found a bug (is.integer() doesn't work when calculating the binwidth, data$x is 'numeric'). I'll try and fix that too.

[tgs]

Ah, the integer/numeric thing seems to be a feature, rather than a bug, of R. OK.

[tgs]

I'm going to open a separate issue for the is.integer thing.

[hadley]

I think that's by design: at this point factors are integers and everything else is numeric. It's definitely weird but it's too late to change now

[tgs]

AH! That makes sense! Ignore #1147 then.

[tgs]

I see that man/stat_bin2d.Rd is deleted in master. Where should documentation go? Is there any chance of getting this change in the next version of ggplot2?

[hadley]

The docs have been merged into geom_bin2d() (since they're basically the same). If you rebase/merge, I can do another round of review and then it should be good to merge.

[tgs]

I've merged, and moved my documentation change to geom_bin2d.Rd. Ready for you to take another look :)

[hadley]

Looks good. Could you please add a bullet to NEWS (copying the existing format) and please rebase (again) now that I've fixed all the build problems?

[tgs]

I rebased on master this morning, it looks like the build passes :)

[tgs]

Woops, let me fix the documentation once more.

[tgs]

OK, fixed.

[hadley]

Thanks!

[tgs]

Thank you too!

[lock]

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