replace multiple values in a column for a single one

Question

I have a rather basic question. I have several values in a column that I would like to replace for a single one, for instance:

a<-data.frame(T=LETTERS[5:20],V=rnorm(16,10,1))

and I would like to change all "E", "S", "T" in T for "AB", so I tried

a[a$T==c("E","S","T")]<-"AB"

and it gives me several warnings, and ends up replacing all to "AB"

I think it has something to do with levels and level's labels but I was not able to replace only some of the values, I would have to re-label each. Sorry for the trouble, and thanks for any help!

Didzis Elferts · Accepted Answer · 2013-06-20 13:14:44Z

7

You can use function recode() from library car to change values also for the factors.

library(car)
a$T<-recode(a$T,"c('E','S','T')='AB'")

If you need to replace different values with different other values then all statements can be written in one function call.

recode(a$T,"c('E','S','T')='AB';c('F','G','H')='CD'")

edited Jun 20, 2013 at 13:14

answered Jun 20, 2013 at 12:45

Didzis Elferts

99.2k17 gold badges275 silver badges207 bronze badges

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

4 Comments

flodel Over a year ago

...but why such an unfriendly interface (not blaming you @Didzis but the author). I read about plyr::mapvalues but I can't test it here on my old R version. I suppose plyr::mapvalues(levels(a$T), c("E","S","T"), "AB") might work if someone can give it a try.

Didzis Elferts Over a year ago

@flodel mapvalues will work with some modification mapvalues(a$T, c("E","S","T"), rep("AB",3))

user1817709 Over a year ago

hello, this is really helpful! Thank you very much. Now I have realized I need to do this for several different groups of T, say (L,M,N). Is it possible to do it for all variables in a single line instead of doing it group by group?

Huanfa Chen Over a year ago

@DidzisElferts Echoing your comment, a$T <- plyr::mapvalues(a$T, c("E","S","T"), rep("AB",3)) is more friendly and readable than car::recode(...)

flodel · Accepted Answer · 2013-06-20 13:39:08Z

4

This would maintain your data structure (a factor like you guessed):

x <- levels(a$T)
levels(a$T) <- ifelse(x %in%  c("E","S","T"), "AB", x)

or

levels(a$T)[levels(a$T) %in%  c("E","S","T")] <- "AB"

Edit: if you have many such replacements, it is a little more complicated but not impossible:

from <- list(c("E","S","T"), c("J", "K", "L"))
to   <- c("AB", "YZ")

find.in.list <- function(x, y) match(TRUE, sapply(y, `%in%`, x = x))
idx.in.list  <- sapply(levels(a$T), find.in.list, from)
levels(a$T)  <- ifelse(is.na(idx.in.list), levels(a$T), to[idx.in.list])

a$T
#  [1] AB F  G  H  I  YZ YZ YZ M  N  O  P  Q  R  AB AB
# Levels: AB F G H I YZ M N O P Q R

edited Jun 20, 2013 at 13:39

answered Jun 20, 2013 at 12:44

flodel

89.3k21 gold badges190 silver badges225 bronze badges

1 Comment

Simon O'Hanlon Over a year ago

+1. this is the right answer, rather than messing around with character class.

statquant · Accepted Answer · 2013-06-20 13:19:35Z

1

Do you really want factors there ??? If not (I think you do not) do options(stringsAsFactors=FALSE) So it is much simpler than that... => a[a$T %in% c("E","S","T"),"T"]<-"AB"

edited Jun 20, 2013 at 13:19

answered Jun 20, 2013 at 13:09

statquant

14.6k23 gold badges98 silver badges172 bronze badges

2 Comments

Didzis Elferts Over a year ago

If T is factor (as in this case) this will give NA values not the AB.

statquant Over a year ago

Ohhhh !!! this is because my options()$stringAsFactors==FALSE, since R.2.12.x R R maintain hash tables for strings I think there is no need for factors in those cases...

Alan Gómez · Accepted Answer · 2024-10-25 01:58:42Z

0

An R Base solution is:

a$T[a$T %in% c("E","S","T")] <- "AB"

answered Oct 25, 2024 at 1:58

Alan Gómez

3881 gold badge4 silver badges11 bronze badges

Collectives™ on Stack Overflow

replace multiple values in a column for a single one

4 Answers 4

4 Comments

1 Comment

2 Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

4 Answers 4

4 Comments

1 Comment

2 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related