[R] How to generate a new factor variable by two other factor variables

[Marc Schwartz]

The easiest way is probably to use interaction():

> df
  factorA factorB
1       0       0
2       0       0
3       1       0
4       0       1
5       1       1


# Note the default separator of '.'
df$factorC <- with(df, interaction(factorA,  factorB))

> df
  factorA factorB factorC
1       0       0     0.0
2       0       0     0.0
3       1       0     1.0
4       0       1     0.1
5       1       1     1.1


If you want numeric values, you can adjust as follows:

df$factorC <- factor(as.numeric(df$factorC) - 1)

> df
  factorA factorB factorC
1       0       0       0
2       0       0       0
3       1       0       1
4       0       1       2
5       1       1       3


The latter step takes advantage of the underlying numeric nature of the
factor levels and subtracts 1, since they are 1 based, not 0 based.

See ?interaction

HTH,

Marc Schwartz


on 11/04/2008 10:10 AM Jorge Ivan Velez wrote:
> Dear Shuguang,
> Here are two ways. Perhaps they are not efficient enough, but the work:
> 
> 
> # Data
> mydata=read.table(textConnection("
> factorA factorB
>    0       0
>    0       0
>    1       0
>    0       1
>    1       1"),header=TRUE)
> closeAllConnections()
> 
> # Option 1
> mydata$factorC=as.factor(
> apply(mydata,1,function(x){
> paste(x,sep="",collapse="")
> }
> ))
> levels(mydata$factorC)<-list("0"="00", "1"="10", "2"="01","3"="11")
> mydata
> 
> # Option 2
> # You'll need to read the data again to see how this option works
> mydata$factorC<-apply(mydata,1,function(x){
> ifelse(sum(x)==0,0,
> ifelse(x[1]==1 & x[2]==0,1,
> ifelse(x[1]==0 & x[2]==1,2,3)))
> }
> )
> 
> mydata
> 
> 
> HTH,
> 
> 
> Jorge
> 
> 
> 
> On Tue, Nov 4, 2008 at 10:29 AM, Shuguang Sun <shuguang at gmail.com> wrote:
> 
>> How to generate a new factor variable by two other factor variables?
>>
>> For example, if I have two factor variables, factorA and factorB,
>> factorA factorB
>>    0       0
>>    0       0
>>    1       0
>>    0       1
>>    1       1
>>
>> Is there a simple way to generate a new 4-levels factor variable as
>>
>> factorC factorA factorB
>>    0      0       0
>>    0      0       0
>>    1      1       0
>>    2      0       1
>>    3      1       1
>>
>> --
>> Shuguang Sun
>> Fudan University, China
>>