Setting `xlim` or `ylim` to values that crop data leads to a confusing warning message

[MilesMcBain]

If data lie outside the bounds set by xlim and ylim the warning issued is:

Warning: Removed # rows containing missing values

This was misinterpreted by a beginner I was helping as an assertion that there were missing data in their dataframe. I think we can make the message more specific in this case, something like:

Warning: Removed # rows containing data outside plot limits.

A reprex:

library(ggplot2)
a <- data.frame(x = 1:3, y = 1:3)
ggplot(a, aes(x = x, y = y)) + 
  geom_point() +
  xlim(0,2) + 
  ylim(0,2)
#> Warning: Removed 1 rows containing missing values (geom_point).

[karawoo]

Setting limits on a scale changes the underlying data and replaces values outside the limits with NAs, hence that message. These are then treated the same as any NAs in the original data, so it may not be easy to tell which NAs were always there and which were introduced by setting the limits, but I can look into it.

Another option that produces the same plot with no warning is:

ggplot(a, aes(x = x, y = y)) +
  geom_point() +
  coord_cartesian(xlim = c(0, 2), ylim = c(0, 2))

But I agree that the message is confusing to beginners, who are also probably more likely to know about xlim and ylim than coord_cartesian.