converting values of a dataframe to numeric (where possible)

df <- data.frame(a=letters[15:17], b=c("21","NA","23"), c=10:12, d=15:17)
# convert to numeric
x <- as.matrix(df)
mode(x) <- "numeric"

cbind(df, RTot=rowSums(x, na.rm=TRUE))

Hi,
I am new to R.
I have a dataframe with many columns some of which contain genuine strings
some numerical values as strings. The dataframe is created by cast so I have
no control over the resulting data type.
I want to attach columns as aggregates of other columns to the dataframe.
Here is the solution that I came up with (after a lot of struggle):

castNum <- function(n) {
   x<-as.numeric(as.character(n))
   if (is.na(x)){
     return(n)
   }else{
     return(x)
   }
}
df <- data.frame(a=letters[15:17], b=c("21","NA","23"), c=10:12, d=15:17)
cbind(df,RTot=rowSums(as.data.frame(lapply(df, function(x)
castNum(x)))[2:4],na.rm=TRUE))

This works, but is full of warnings and looks extremely ugly.
Could you direct me how to achieve the same result in a more elegant way?

Thx.
Bal?zs



--
View this message in context: http://www.nabble.com/converting-values-of-a-dataframe-to-numeric-%28where-possible%29-tp19279139p19279139.html
Sent from the R help mailing list archive at Nabble.com.