CODEWITHSUNDEEP
r
Mahi
Mahi

Reputation: 59

If () statement in R

I am fairly new in R and have question on if () statement in R. Any help is greatly appreciated. My data look like as listed below

Var1    var2 
01         A
02         D
03         E
04         K

I need to create 2 additional variables (Var3 and Var4) based on var1. In SAS I would create var3 and var4 in following way:

if var1 = 01 then do; var3 = "New York";   var4= "NYC"; end; else;
if var1 = 02 then do; var3 = "Illinois";   var4= "ORD"; end; else;
if var1 = 03 then do; var3 = "Texas";      var4= "DFW"; end; else;
if var1 = 04 then do; var3 = "New Jersey"; var4= "EWR"; end; else;

I want to convert above if statements into R program. Any help is greatly appreciated.

Upvotes: 0

Views: 713

Answers (3)

thelatemail
thelatemail

Reputation: 93813

The basic R rule I impose on myself is to store data that you might have to call upon in an appropriate structure (e.g. - vectors for a series of values, lists for related series of values of unequal lengths, data.frames for related series of equal lengths etc etc).

So an appropriate structure in this case is a data.frame for related series of the same length. E.g.:

srcs <- data.frame(
  Var1=c("01","02","03","04"),
  state=c("New York", "Illinois", "Texas", "New Jersey"),
  code=c("NYC", "ORD", "DFW", "EWR")
)

Then use that store of basic information to add new variables as appropriate, e.g.:

dat <- cbind(dat, srcs[c("state","code")][match(dat$Var1,srcs$Var1),] )

#  Var1 var2      state code
#1   01    A   New York  NYC
#2   02    D   Illinois  ORD
#3   03    E      Texas  DFW
#4   04    K New Jersey  EWR

This turns a whole bunch of if/ else statements into a single call, and avoids the need for ugly looping operations.

Upvotes: 2

Metrics
Metrics

Reputation: 15458

Another solution is to use ifelse (assuming that you have only four values for var1). 

mydata$var3<- with (mydata, ifelse(var1 =="01", "New York" ,
              ifelse (var1 == "02", "Illinois",
              ifelse (var1 == "03","Texas" ,"New Jersey" ))))

mydata$var4<-with (mydata, ifelse (var1 =="01", "NYC",
        ifelse (var1 == "02","ORD",
      ifelse (var1 == "03", "DFW","EWR"))))

Upvotes: 0

IRTFM
IRTFM

Reputation: 263331

var3 <- c("New York", "Illinois", "Texas", "New Jersey")[
                 match(var1, c("01","02","03","04")]
var4 <- c("NYC", "ORD", "DFW", "EWR")[
                 match(var1, c("01","02","03","04")]

Generally the items you will be working with are dataframes rather than the Sets you have in SAS-world. So you would have a data-object name, say dat and you might execute:

dat$var4 <- with( dat, c("NYC", "ORD", "DFW", "EWR")[
                 match(var1, c("01","02","03","04")] )

Upvotes: 5

Related Questions