Deparse substitute assign with list elements

Thanks, Bill, I should have googled more carefully:


http://stackoverflow.com/questions/9561053/assign-values-to-a-list-element-in-r

So, remove

        assign(nm, tmp, parent.frame())

and add

        txt <- paste( nm, '<-', tmp, sep='' )
        eval( parse(text=txt), parent.frame() )

in `foo()` will do the trick.

Bests
S?ren

On 15.05.2015, at 00:02, William Dunlap <wdunlap at tibco.com> wrote:

You could use a 'replacement function' named 'bar<-', whose last argument
is called 'value', and use bar(variable) <- newValue where you currently
use foo(variable, newValue).

bar <- function(x) {
    x + 3
}
`bar<-` <- function(x, value) {
    bar(value)
}

a <- NA
bar(a) <- 4
a
# [1] 7
b <- list(NA, NA)
bar(b[[1]]) <- 4
b
#[[1]]
#[1] 7
#
#[[2]]
#[1] NA


Bill Dunlap
TIBCO Software
wdunlap tibco.com

On Thu, May 14, 2015 at 11:28 AM, <soeren.vogel at posteo.ch> wrote:
Hello,

When I use function `foo` with list elements (example 2), it defines a

new object named `b[[1]]`, which is not what I want.  How can I change the
function code to show the desired behaviour for all data structures passed
to the function?  Or is there a more appropriate way to sort of "pass by
references" in a function?

Thanks
S?ren

<src>
bar <- function(x) {
        return( x + 3 )
}

foo <- function(x, value) {
        nm <- deparse(substitute(x))
        tmp <- bar( value )
        assign(nm, tmp, parent.frame())
}

# 1)
a <- NA
foo(a, 4)
a # 7, fine :-)

# 2)
b <- list(NA, NA)
foo(b[[1]], 4) # the first list item should be 7
b # wanted 7 but still list with two NAs :-(
</src>