On Wed, Jun 24, 2009 at 12:34 PM, Mark Na<mtb...@gmail.com> wrote:
> The problem is that after running the ifelse statement, data$SOCIAL_STATUS
> is converted from a factor to a character.
> Is there some way I can avoid this conversion?
I'm afraid that ifelse has very bizarre semantics when the yes and no
arguments don't have the same, atomic vector, type.
The quick workaround for the bizarre semantics (though it can have a
significant efficiency cost) is this:
unlist( ifelse ( condition, as.list( yes ), as.list( no ) ) )
(This isn't perfect, either, but...)
Take a look at the man page for details and the warning:
The mode of the result may depend on the value of 'test', and the
class attribute of the result is taken from 'test' and may be
inappropriate for the values selected from 'yes' and 'no'.
Some consequences of the definition of ifelse are:
Even if the classes of the yes and no arguments are identical, the
result does not necessarily have that class:
ifelse(TRUE,as.raw(4),as.raw(5)) => error
ifelse(TRUE,factor('x'),factor('x')) => 1 (integer)
dates <- as.POSIXct(c('1990-1-1','2000-1-1'))
ifelse(c(TRUE,FALSE),dates,dates) => 631170000 946702800 (double)
ifelse(c(TRUE,FALSE),factor(c('x','y')),factor(c('y','x'))) => 1 1
If they have different classes, things get stranger:
ifelse(c(TRUE,FALSE),c("a","b"),factor(c("c","d"))) => "a" "2"
ifelse(c(TRUE,FALSE),list(1,2),as.raw(4))
[[1]]
[1] 1
[[2]]
[1] 04
Result is order-dependent:
ifelse(c(TRUE,FALSE),as.raw(4),list(1,2))
Error in ans[test & !nas] <- rep(yes, length.out = length(ans))[test & :
incompatible types (from raw to logical) in subassignment type fix
Welcome to R!
-s
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
