[BioC] reducing RangedData

[Janet Young]

Hi Patrick,

I second Robert's thanks - reduce on RangedData is really useful.  

I have an additional suggestion (you might already be on the case with this 
one).  I was trying to reduce a RangedData object that didn't have any 
strand/score/name, and kept getting an error 
(Error in endoapply(x[, by], FUN) : 'FUN' did not produce an endomorphism).
I got around that by just adding a fake name (or strand or whatever) that is
identical for all objects.  

It would be even nicer if reduce could handle bare bones RangedData objects 
without needing this workaround.

Here is some code to show you what I mean, following your example:

suppressMessages(library(IRanges))
sta <- c(1, 2, 5, 6, 2, 3, 4, 5)
end <- c(3, 4, 5, 7, 2, 4, 6, 7)
str <- rep(c("+", "+", "-", "-"), 2)
chr <- rep(c("chr1", "chr2"), each = 4)
rd <- RangedData(IRanges(start=sta, end=end), strand=str, space=chr)
reduce(rd)   ####works fine

rd2 <- RangedData(IRanges(start=sta, end=end), space=chr)
reduce(rd2)
####Error in endoapply(x[, by], FUN) : 'FUN' did not produce an endomorphism

rd3 <- RangedData(IRanges(start=sta, end=end), fakename=rep("hello",8),space=chr)
reduce(rd3)   ####works fine

thanks very much,

Janet Young