On Tue, 2 Mar 2004 16:56:58 +0100, you wrote:
>I wanted to do this, but in R when you have a single object as argument
>you can
>retrieve its name by the deparse(substitute())-trick, but if the
>function is just
>given as part of a list element there is no way (thai I know of) to get
>at its
>name. You can of course do the deparse(substitute())-trick, and embark
>on major
>text-processing to get you the name.
I was just thinking of the first case, where you had something like
margins(A, FUN=mean)
but I think you can handle the more complicated case using
substitute(FUN), i.e. and using this function, which recursively goes
through a list, and adds a name matching the value when the value is
simple.
add.names <- function(thelist) {
n <- names(thelist)
if (is.null(n)) n <- rep("", length(thelist))
for (i in seq(along=thelist)[-1]) {
if (!is.call(thelist[[i]])) {
if (n[i] == "") n[i] <- as.character(thelist[[i]])
} else if (as.character(thelist[[i]][[1]]) == "list")
thelist[[i]] <- add.names(thelist[[i]])
}
names(thelist) <- n
thelist
}
I will add your function to the new "stats" package which is in the
base R distribution (the same package holding ftable). I think it
would be better to name it "add.margins", since "margins" sounds like
a function to extract margins. What do you think of that change?
Duncan Murdoch
______________________________________________
[EMAIL PROTECTED] mailing list
https://www.stat.math.ethz.ch/mailman/listinfo/r-devel
