bug: edge case for histogram

[lixun910]

using the NYC data, I created a custom break with six categories for kids2000. That’s illustrated in the updated mapping chapter. When I use this category in a histogram for kids2000, all is well. However, when I use
it in a histogram for kids2009, it gives a very weird result on the axes and the statistics.
As it happens, kids2009 does not have any observations over 50, which is the highest
breakpoint in custom1, the custom category. The histogram drawing code doesn’t seem
to know how to deal with that. See also, the strange values for the range in the last bin
in the statistics part, going from 50 to 48.1!