add column values based on other columns in data frame using for and if

Question

I have a dataframe like this:

     id        adit     diag1   diag2       
      2       3         4230    2234        
      3       5         3345    4456        
      4       6         4567    4467

I would like to add other 2 columns, dse1 and dse2 using the pseudo-code below:

if diag1 contains 4230 then dse1 = 1 else dse1 = 0

if diag2 contains 4567 then dse2 =1  else dse2 = 0

I used this:

for (i in 1 : nrow(dse)){
  for (j in 3: ncol(dse)){
     if dse[i,j] %in% ("4320"){dse$dse1 = 1}
        else{dse$dse1 = 0}
    if dse[i,j] %in% ("4567"){dse$dse2 = 1}
        else{dse$dse2 = 0} 
  }
}

But these do not work.

All the answers below are the correct route. However, if you are using if/else then the else needs to be on the same line as the closing }` from your if statement. — Justin
– Justin, Commented Jan 7, 2013 at 18:53
for (i in 1 : nrow(dse)){ for (j in 3: ncol(dse)){ if ( dse[i,j] == 4230) { dse$dse1 = 1 } else{ dse$dse1 = 0 } if (dse[i,j] %in% (4567)) {dse$dse2 = 1} else{dse$dse2 = 0} } } I change the codes as above, there is no error, But the result is not correct. Why? THnaks — user1582755
– user1582755, Commented Jan 8, 2013 at 19:15
Not sure what to make of your comment. You can edit your question as needed. Also, if any of the answers below has solved the problem, please mark it as answered by clicking the green check mark by the question so we all know its been resolved. — Justin
– Justin, Commented Jan 8, 2013 at 19:18
I know the problems, the dse$dse1 should be dse$dse[i], and dse$dse2 should be dse$dse2[i]. Thanks for your help and answers — user1582755
– user1582755, Commented Jan 8, 2013 at 19:32

Paul Hiemstra · Accepted Answer · 2013-01-07 18:54:28Z

25

No need to use a loop, just use ifelse, for example

dse = within(dse, {
    dse1 = ifelse(diag1 == 4230, 1, 0)
    dse2 = ifelse(diag2 == 4567, 1, 0)
 })

edited Jan 7, 2013 at 18:54

answered Jan 7, 2013 at 18:48

Paul Hiemstra

61.2k12 gold badges146 silver badges151 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

user1582755 Over a year ago

Thanks. The dse1 maybe has many diagnoses, this is, if dse[i,j] %in% ("2345","3456","5678") {dse1 =1 ). And the if codes have any woring? how to correct them

Paul Hiemstra Over a year ago

I am not sure what you want, please add more information to your question.

thelatemail Over a year ago

@user1582755 - To adjust the code for multiple matches I think you would just replace the corresponding lines above like: dse1 = ifelse(diag1 %in% c(2345,3456,5678), 1, 0)

user1582755 Over a year ago

dse1 = 1 if diag1 or diag2, etc in('224','334',333' etc) ; dse2 = 1 if diag1 or diag2 or diag3(if having) in('444','435',etc)

Tyler Rinker · Accepted Answer · 2013-01-07 18:47:42Z

13

Don't use the if/else. Go vectorized as in:

dat$dse1 <- as.numeric(dat$diag1 == 4230)
dat$dse2 <- as.numeric(dat$diag2 == 4567)

answered Jan 7, 2013 at 18:47

Tyler Rinker

111k74 gold badges335 silver badges535 bronze badges

2 Comments

Paul Hiemstra Over a year ago

+1 nice answer, if feel however that using ifelse is a little more transparent. Also, you can have other vales than 0 and 1.

Rich Scriven Over a year ago

This is definitely the best answer. ifelse(cond, 1, 0) should never be used.

Theodore Lytras · Accepted Answer · 2013-01-07 18:50:41Z

5

Like this:

dse$dse1<-0
dse$dse2<-0
dse$dse1[dse$diag1==4230]<-1
dse$dse2[dse$diag2==4567]<-1

Please get yourself a good R tutorial (such as this) and read all about index vectors.

answered Jan 7, 2013 at 18:50

Theodore Lytras

3,9631 gold badge21 silver badges25 bronze badges

1 Comment

user1582755 Over a year ago

Thanks for all answers. If we change the statements like these if diag1 OR diag2 contains 4230 then dse1 = 1 else dse1 = 0 if diag2 or diag1 contains 4567 then dse2 =1 else dse2 = 0 and there are many diags in which there are many codes, so there are many dses. What are the neat answers

Sven Hohenstein · Accepted Answer · 2017-01-15 07:49:58Z

4

You can use transform:

transform(dse, dse1 = as.numeric(diag1 == 4230),
               dse2 = as.numeric(diag2 == 4567))

edited Jan 15, 2017 at 7:49

answered Jan 7, 2013 at 18:49

Sven Hohenstein

82k17 gold badges150 silver badges173 bronze badges

1 Comment

user1582755 Over a year ago

Thanks. The dse1 maybe has many diagnoses, this is, if dse[i,j] %in% ("2345","3456","5678") {dse1 =1 ).

IndPythCoder · Accepted Answer · 2018-06-19 04:44:09Z

0

you can also use:

ifelse():

dat <- data.frame(id = c(2,3,4), adit = c(3,5,6),diag1 = c(4230,3345,4567), diag2 =            c(2234,4567,4467))
dat$dse1 <- ifelse(dat$diag1 == 4230,1,0)
dat$dse2 <- ifelse(dat$diag2 == 4567,1,0)
dat

answered Jun 19, 2018 at 4:44

IndPythCoder

7631 gold badge7 silver badges11 bronze badges

Comments

pedrostrusso · Accepted Answer · 2020-01-15 15:50:09Z

0

A solution using tidyverse:

x = data.frame(id = c(2, 3, 4), 
               adit=c(3, 5, 6), 
               diag1=c(4230, 3345, 4567), 
               diag2=c(2234, 4456, 4467))

x %>% mutate(dse1 = if_else(diag1 == 4230, 1, 0), 
             dse2 = if_else(diag2 == 4567, 1, 0))

answered Jan 15, 2020 at 15:50

pedrostrusso

3984 silver badges12 bronze badges

Collectives™ on Stack Overflow

add column values based on other columns in data frame using for and if

6 Answers 6

4 Comments

2 Comments

1 Comment

1 Comment

Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

6 Answers 6

4 Comments

2 Comments

1 Comment

1 Comment

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related