[R] sort table

[Marc Schwartz]

On Sat, 2005-11-05 at 11:44 -0600, Marc Schwartz wrote:
> On Sat, 2005-11-05 at 18:04 +0100, Claus Atzenbeck wrote:
> > Hi,
> > 
> > I have a data frame named "questions" that I use to get a subset and
> > then a table:
> > 
> > failcondonly <- subset(questions, errorreason=="condition")
> > failcondtab <- table(failcondonly$type, failcondonly$qid, exclude=c("ORGA", "skipped"))
> > 
> > The failcondtab looks like this:
> > 
> >         6 11 12 13 14 15 17 26 30 31 39 41
> > gave up 0  1  1  1  1  0  0  0  0  8  0  2
> > wrong   3  3  7  3  3  1  4  1  3  4  1  2
> > 
> > I want to sort this table according to the *sum* of the values in "gave
> > up" and "wrong" in reverse order. The highest value should be first.
> > First would be column "31" (value 8+4=12), then column "12" (value
> > 1+7=8), etc.
> > 
> > sort(failcondtab) does not help, since it sorts the column names, not
> > the sum of the values of each column.
> > 
> > Thanks for any hint.
> > Claus
> 
> 
> Something like the following should work:
> 
> > failcondtab[, rev(order(colSums(failcondtab)))]
>         31 12 41 17 14 13 11 30 6 39 26 15
> gave up  8  1  2  0  1  1  1  0 0  0  0  0
> wrong    4  7  2  4  3  3  3  3 3  1  1  1
> 
> 
> See ?order, ?rev and ?colSums.

Or another, better variation, would be:

> failcondtab[, order(colSums(failcondtab), decreasing = TRUE)]
        31 12 11 13 14 17 41 6 30 15 26 39
gave up  8  1  1  1  1  0  2 0  0  0  0  0
wrong    4  7  3  3  3  4  2 3  3  1  1  1

which removes the requirement to use rev().

Marc