On 12 Mar 2012, at 12:47 , S Ellison wrote: > Yes, to the extent that the default barplot plots the height of the bar so > far as the sum of teh values so far, starting at teh first. For your first > vector, no problem; for your second, the highest value is undefiuned, for the > third, the sum is undefined after the second value (an NA) and so on. > > Try adding 'beside=TRUE to the barplots, as in > barplot(d3, beside=TRUE) > and you will see all the known values plotted as you;d expect.
That makes sense, but since I do want a stacked bar plot, I'll need to change the NAs to 0 (which of course I've already done). This should be made clear in the documentation, no? It's possible that barplot could do something like a na.rm=T internally and avoid this problem, but it doesn't, so NA is deadly in stacked plots. To be honest, if I hadn't scaled all my bars to 1 to show percentages, I wouldn't have noticed how some were leaving out a small category or two. Thanks for the help. John Muccigrosso ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.