[R] Conflating categories

[Duncan Murdoch]

On 12/14/2007 7:39 AM, Peter Paul wrote:
> 
> Hi,
> 
> I think this is a pretty basic question. I still couldn#t find the answer to 
> it, though.
> 
> I have some data loaded into R, which looks like this:
> 
>> data()
> ...
> 38358     Advice      Article
> 38359     Advice      Article
> 38360    GeneralInfo  List
> 38361    GeneralInfo      Article
> 38362     Purchase   Paragraphs
> 38363     Purchase         List
> 38364     Purchase   Paragraphs
> ...
> 
> 
> I now would like to edit the data in such a way that every "Advice" is 
> changed into "GeneralInfo".
> Is there some easy way how I can do this?

This is a little tricky, because R has probably converted those columns 
into factors.  Let's assume that column 2 is named Type and the dataset 
is named dat.

Then you replace it as

dat$Type[dat$Type == "Advice"] <- "GeneralInfo"

You'd think the following would work:

dat$Type <- ifelse(dat$Type == "Advice", "GeneralInfo", dat$Type)

but it doesn't, because ifelse() loses the factor levels.  Factors are 
one of the most error-prone features of the S language.

Duncan Murdoch