[R] Recode function car package erases previous values

[Pancho Mulongeni]

-----Original Message-----
From: arun [mailto:smartpink111 at yahoo.com] 
Sent: Wednesday, October 24, 2012 3:31 PM
To: Pancho Mulongeni
Cc: R help
Subject: Re: [R] Recode function car package erases previous values

Hi,
May be this helps:
set.seed(1)
dat1<-data.frame(B20_C1=c(NA,sample(0:1,4,replace=TRUE),NA),B20_C2=c(NA,sample(0:1,3,replace=TRUE),sample(0:1,2,replace=TRUE)),nrB20C=c(1,NA,NA,NA,NA,NA))
dat1
#  B20_C1 B20_C2 nrB20C
#1     NA     NA      1
#2      0      0     NA
#3      0      1     NA
#4      0      0     NA
#5      1      1     NA
#6     NA      1     NA
 dat1$pharm<-ifelse(apply(dat1,1,function(x) all(is.na(x[1:2]))),"no response",ifelse(dat1[,1]==1,"Yes","No"))
dat2<-within(dat1,{pharm<-factor(pharm)})
 levels(dat2$pharm)
#[1] "No"          "no response" "Yes"
 dat2
#  B20_C1 B20_C2 nrB20C       pharm
#1     NA     NA      1 no response
#2      0      0     NA          No
#3      0      1     NA          No
#4      0      0     NA          No
#5      1      1     NA         Yes
#6     NA      1     NA        <NA>

table(dat2$pharm)

#         No no response         Yes
 #         3           1           1 

A.K.

Thank you, once I understand what you did I will use it, for now, I use a workaround
1. First I use the recode function
pharm<-as.factor(recode(B20_C1,"1='Yes';0='No'"))
2. Now I just subset
pharm[nr.B20C==1]<-'no resp'
and this gives me the desired output
     No     Yes no resp
    716       7       6 


----- Original Message -----
From: Pancho Mulongeni <p.mulongeni at namibia.pharmaccess.org>
To: "r-help at r-project.org" <r-help at r-project.org>
Cc: 
Sent: Wednesday, October 24, 2012 5:17 AM
Subject: [R] Recode function car package erases previous values

Hi all,
I am attempting to create a new variable based on values of other variables. The variable is called pharm. It basically takes the numeric code of 1 as yes and 0 to be No from the variable B20_C1 (a question on a survey). However, I would also like to have a level for non-respondents and these are captured in the variable nr.B20C, which is a 1 when there is a non-response on the whole group of variable B20_C1 to B20_C5. So ultimately the variable pharmacy will have three levels, Yes No and no resp.
See below what happens

> pharm<-as.factor(recode(B20_C1,"1='Yes';0='No'"))
> table(pharm)
pharm
No Yes
716   7 
> levels(pharm)<-c('No','Yes','no resp')
> table(pharm)
pharm
     No     Yes no resp
    716       7       0 
> pharm<-as.factor(recode(nr.B20C,"1='no resp'"))
> table(pharm)
pharm
      0 no resp
    723       6 
>
The recode variable just cannot seem to 'remember' I had just recoded 7 values to 'Yes' and 716 to be 'No'  and instead it assigns the level '0' which comes from nr.B20C (it has values 0 or 1). 
This inconvenient as I would like to have ultimately the following table pharm
     No     Yes no resp
    716       7       6 (FROM nr.B20C where row has value 1).

Background. The variable pharm assess where you used the pharmacy to get your contraception.

Pancho Mulongeni
Research Assistant
PharmAccess Foundation
1 Fouché Street
Windhoek West
Windhoek
Namibia
 
Tel:   +264 61 419 000
Fax:  +264 61 419 001/2
Mob: +264 81 4456 286

______________________________________________
R-help at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.