Skip to content

Recoding variables in R

3 messages · Lisa van der Burgh, William Dunlap, Rui Barradas

Lisa van der Burgh
# 
Hi all,


I have a very general question and I think for you maybe very easy, but I am not able to solve it.

I have a dataset and that dataset contains the variable Anxiolytics. This variable is coded as 0, 1, 2. The variable looks as follows:
0        1     2       NA's
1102   0     20    440


You can see that the variable is coded as 0, 1, 2, but group 1 is 'empty'. So I want to remove this group, such that I get:
0          2       NA's
1102     20      440

And then I want to recode the variable in a way that 2 becomes 1, such as:
0          1       NA's
1102     20      440

Can you help me with the code for doing this?

Thank you in advance!
Best, Lisa
William Dunlap
# 
It looks like your data has class "factor".  If you call factor() on a
factor variable, without supplying an explicit 'levels' argument it
produces a new factor variable without any levels not present in the input
factor.  E.g.,
D    C    B    A NA's
   2    0    1    1    2
D    B    A NA's
   2    1    1    2

You can use levels(fNew) <- c("delta","beta","alpha") to give the levels
new names.

I think that renaming "0", "2", "3"  to "0","1","2" will lead to
confusion.  The levels
are just names, so give them names that mean something to a human, like
"control",
"streptomycin (2 mg/kg)", etc.


Bill Dunlap
TIBCO Software
wdunlap tibco.com

On Wed, May 23, 2018 at 3:14 AM, Lisa van der Burgh <lisavdburgh at hotmail.com

  
  


  


      

    

    

      
      

        


  

        

      Rui Barradas
    

    

      
      #
    

  


  

      
Hello,

See if this inspires you.


set.seed(1962)
DF <- data.frame(Anxiolytics = factor(sample(c(0, 2, NA), 1000, TRUE), 
levels = 0:2))
summary(DF$Anxiolytics)

DF$Anxiolytics <- droplevels(DF$Anxiolytics)
summary(DF$Anxiolytics)

DF$Anxiolytics <- factor(DF$Anxiolytics, labels = 0:1)
summary(DF$Anxiolytics)


Hope this helps,

Rui Barradas

        
On 5/23/2018 11:14 AM, Lisa van der Burgh wrote: