R help - Dec 2002 - histogram bug?: type='count', unequal breaks

I ask for a histogram of counts with slightly uneven breaks but
histogram() ignores the 'type' argument and provides density scaling
instead.

   x = sample(1:3, 100, replace=TRUE)
   histogram( ~ x, breaks=c(0,1.5,2.5,3.5), type='count')

My real application has time data with month boundaries for the breaks
and I DO want counts on the y-axis.

A work-around is not difficult but shouldn't histogram honor the
type argument when given?

 Scott Vander Wiel
 Statistics and Data Mining Research
 Bell Laboratories, Lucent Technologies

On Friday 13 December 2002 12:43 pm, scottyv at bell-labs.com
wrote:
> I ask for a histogram of counts with slightly uneven breaks but
> histogram() ignores the 'type' argument and provides density
scaling
> instead.
>
>    x = sample(1:3, 100, replace=TRUE)
>    histogram( ~ x, breaks=c(0,1.5,2.5,3.5), type='count')
>
> My real application has time data with month boundaries for the breaks
> and I DO want counts on the y-axis.
>
> A work-around is not difficult but shouldn't histogram honor the
> type argument when given?
Yes, it should, and it doesn't because of a bug in the released version of 
lattice. This should be fixed in the next release (hopefully sometime next 
week).

Deepayan