[R] Reshape with var as fun.aggregate

[locklin.jason at gmail.com]

I used to be able to use variance for fun.aggregate, but since I 
upgraded from R 2.6 to 2.7 (Ubuntu hardy to intrepid), it no longer 
works.

> library("reshape")
> data("french_fries")
> ffm <- melt(french_fries, id=1:4, na.rm = TRUE)
> cast(ffm, rep ~ treatment, length)
  rep   1   2   3
1   1 579 578 575
2   2 580 579 580
> cast(ffm, rep ~ treatment, mean)
  rep        1        2        3
1   1 3.207772 3.177336 3.024522
2   2 3.181207 3.115544 3.277759
> cast(ffm, rep ~ treatment, var)
Error in fun.aggregate(data$value[0]) : 'x' is empty

Any ideas?