mutate/transform in R dplyr (Pass custom function)

Question

I am using the new package , dplyr and facing some difficulties.

mutate(df,isOdd=digit%%2) or transform(df,isOdd=digit%%2)

Both of which work perfectly.

I am asking a question on passing custom method.

IsItOdd <- function(x) { if(x%%2==0) result<-"even" else result<-"odd" return(result) } transform(df,isOdd=IsItOdd(digit))

This doesnt work because the whole column of all the digit is passed to the function. Is there a way to make this work by just passing that one cell to the function instead of the whole column ?

transform(df,isOdd=vapply(digit, IsItOdd, "")) or transform(df,isOdd=Vectorize(IsItOdd)(digit))? — lukeA
– lukeA, Commented Feb 18, 2015 at 8:55

Mathew Hall · Accepted Answer · 2015-02-18 08:55:10Z

15

With transform your function has to operate on the vector. You can use ifelse instead, which works on vectors:

 isOdd <- function(x){ ifelse(x %% 2 == 0, "even", "odd") }

Alternatively you can apply the function to every value in the column with one of the apply functions:

 isOdd <- function(x){ sapply(x, function(x){ if(x %% 2 == 0){ return("even") }else{ return("odd") } })}

answered Feb 18, 2015 at 8:55

Mathew Hall

1,00410 silver badges19 bronze badges

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

4 Comments

Robert McDonald Over a year ago

Can you explain why with transform/mutate the function operates on the entire vector rather than on the individual values? This isn't true with base functions (e.g. sin() or log()).

Will Cornwell Over a year ago

@rmcd I'm very interested in your question, maybe worth another thread? I guess dplyr::mutate is parsing the function and trying to guess whether to pass in a single value or the whole vector?

Will Cornwell Over a year ago

btw you can force the proper behavior by using dplyr::rowwise but it's still an interesting question

Robert McDonald Over a year ago

@WillCornwell Thanks for the response and the pointer to rowwise, which I was't aware of. I've posted a question here stackoverflow.com/questions/49967559/…

Extrapolator · Accepted Answer · 2015-11-16 10:07:56Z

I think you could also use group_by() to tease apart the rows by unique values and subsequently do your computation, like so:

df %>% group_by(digit) %>% mutate(isOdd = IsItOdd(digit))

Jorge Ramirez · Accepted Answer · 2018-07-11 19:03:12Z

You do not need to use mutate, you can do it in R base or in purr

get_rango_edad <- function(x) { if (x <= 25) { return("18-25") } else{ return("26+") } } encuestas$rango_edad <- map_chr(encuestas$edad,get_rango_edad)

or

encuestas$rango_edad <- sapply(encuestas$edad,get_rango_edad)

Collectives™ on Stack Overflow

mutate/transform in R dplyr (Pass custom function)

3 Answers 3

4 Comments

Comments

Comments

Linked

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

4 Comments

Comments

Comments

Linked

Related