[R] change value in one cell

[Rolf Turner]

On 11/06/11 18:14, jour4life wrote:
> That's great! But, I guess I should have provided a better example for
> my problem. The reason being that I did not consider strings. For
> instance, in my case, I'm actually trying to add a value to a data frame
> that looks like this:
>
> Obs     X
> 0     NA
> 1     "01 001"
> 2     "01 002"
> 3     "01 003"
>
> And I am actually trying to get that specific "NA" value and convert it
> to "00 000" so it can look like this:
>
> Obs     X
> 0     "00 000"
> 1     "01 001"
> 2     "01 002"
> 3     "01 003"
>
> When I write the code you provided, I get this result:
>
> x[1,"X"]<-"00 000"
> Warning message:
> In `[<-.factor`(`*tmp*`, iseq, value = c(NA, 3L, 4L, 5L, 6L, 7L,  :
>
> invalid factor level, NAs generated
>
> I am wondering what I am doing wrong and how to solve this problem.
>
> Thanks for all the help!

The error is clear enough, isn't it?  "00 000" is not one of the levels 
of the
"X" column of your data frame; this column is a factor.  Probably because
of the "stringsAsFactors = TRUE" default in options().

Since you appear to have been unaware of the factor nature of "X", 
presumably
you don't really want it to be a factor.  If this is the case execute

     x[,"X"] <- as.character(x[,"X"])

and then your reassignment of the [1,"X"] entry of "x" will work.

If you do want "X" to be a factor you could:

     (a) execute x[,"X"] <- factor(x[,"X"]) *after* doing the 
reassignment, or

     (b) execute levels(x[,"X"]) <- c("00 000",levels(x[,"X"])) *before* 
doing
     the reassignment.

Learn more about how R works.  In particular learn about factors; they are
important and useful.

     cheers,

         Rolf Turner