Changing column names of a data frame

Question

I have a data frame called "newprice" (see below) and I want to change the column names in my program in R.

> newprice
   Chang.  Chang.   Chang.
1     100       36      136
2     120      -33       87
3     150       14      164

In fact this is what am doing:

names(newprice)[1]<-paste("premium")
names(newprice)[2]<-paste("change")
names(newprice)[3]<-paste("newprice")

I have not put this in a loop because I want each column name to be different as you see.

When I paste my program into R console this is the output it gives me:

> names(newprice)[1]<-paste(“premium”)
Error: unexpected input in "names(newprice)[1]<-paste(“"
> names(newprice)[2]<-paste(“change”)
Error: unexpected input in "names(newprice)[2]<-paste(“"
> names(newprice)[3]<-paste(“newpremium”)
Error: unexpected input in "names(newprice)[3]<-paste(“"

I have equally tried using the c() function-for example c("premium"), instead of the paste() function, but to no avail.

Could someone help me to figure this out?

If Dirk's answer works then the problem was that you were working with a matrix rather than with a dataframe. You can check this with either `is.matrix` or `str`. — IRTFM, May 21 '11 at 12:17
See this answer on dplyr::rename http://stackoverflow.com/a/26146202/1831980 — Rasmus Larsen, Feb 08 '16 at 11:53
Your error has nothing to do with the quality of your code. You're just using the wrong symbol. This “ is not recognized by R, use " instead. I know they may look the same. Look close: “ ". That's it. — Edo, Nov 30 '19 at 01:44
Several answers below using a hard-coded position, for instance, `2` in `colnames(X)[2]`. This is usually a not good practice because it is sensitive to data change. What if you add another column before this specific column to your data? Instead, try something like the answer provided by **Hagos**. — Jia Gao, May 25 '22 at 00:04

Dirk Eddelbuettel · Answer 1 · 2016-07-09T11:28:30.953

685

Use the colnames() function:

R> X <- data.frame(bad=1:3, worse=rnorm(3))
R> X
  bad     worse
1   1 -2.440467
2   2  1.320113
3   3 -0.306639
R> colnames(X) <- c("good", "better")
R> X
  good    better
1    1 -2.440467
2    2  1.320113
3    3 -0.306639

You can also subset:

R> colnames(X)[2] <- "superduper"

edited Jul 09 '16 at 11:28

answered May 21 '11 at 11:45

Dirk Eddelbuettel

360,940
56
644
725

16

@Dirk Why not using names() instead of colnames()? – Antoine Lizée Oct 10 '13 at 06:40
9

Great! You can also subset multiple columns at once (useful on big data frames). `colnames(X)[c(1,2)] <- c("good", "better")` – metakermit Nov 13 '13 at 12:07
8

Try `setnames()` in the `data.table` package. Use something like `setnames(DT,"b","B")` or `setnames(DT,c("a","E"),c("A","F"))` – dwstu Aug 11 '14 at 17:23
Weirdly, after setting the column names of the data frame `q1`, trying to mutate the data frame using `dplyr` as in `q1 <- q1 %>% mutate(rel_count = count / 482462)` results in the error `Error in mutate_impl(.data, dots) : unknown column 'days'` (where `days` is a new name given to the column). This is really frustrating. – David Tonhofer Feb 07 '17 at 21:19

score 232 · Answer 2 · edited Jun 17 '15 at 02:18

232

I use this:

colnames(dataframe)[which(names(dataframe) == "columnName")] <- "newColumnName"

edited Jun 17 '15 at 02:18

Rorschach

31,301
5
78
129

answered Mar 17 '14 at 18:00

Matheus Abreu

3,357
1
18
21

11

Thank you. I think this is somehow annoying with R: Why is it so difficult to change the column name if you do not want to use the index number but the old name :( – Arne Mar 18 '14 at 14:41
12

This method has the advantage that you do not have to worry about the position of the column, as long as you know its original name. I think this is the preferred method as you may - later - make changes to the code that change the position of the column you want to rename. – Paulo S. Abreu Dec 21 '15 at 19:00
1

Can also use `data.table::setnames(dataframe,'Old','New')` – kakarot Jan 03 '22 at 19:23

Joshua Ulrich · Answer 3 · 2011-05-21T12:48:20.170

The error is caused by the "smart-quotes" (or whatever they're called). The lesson here is, "don't write your code in an 'editor' that converts quotes to smart-quotes".

names(newprice)[1]<-paste(“premium”)  # error
names(newprice)[1]<-paste("premium")  # works

Also, you don't need paste("premium") (the call to paste is redundant) and it's a good idea to put spaces around <- to avoid confusion (e.g. x <- -10; if(x<-3) "hi" else "bye"; x).

score 75 · Answer 4 · edited Jan 13 '21 at 09:55

75

Try:

names(newprice)[1] <- "premium"

edited Jan 13 '21 at 09:55

zx8754

52,746
12
114
209

answered May 21 '11 at 11:36

Jamie

5,994
1
18
15

score 55 · Answer 5 · edited Nov 06 '15 at 14:46

The new recommended way to do this is to use the setNames function. See ?setNames. Since this creates a new copy of the data.frame, be sure to assign the result to the original data.frame, if that is your intention.

data_frame <- setNames(data_frame, c("premium","change","newprice"))

Newer versions of R will give you warning if you use colnames in some of the ways suggested by earlier answers.

If this were a data.table instead, you could use the data.table function setnames, which can modify specific column names or a single column name by reference:

setnames(data_table, "old-name", "new-name")

Desta Haileselassie Hagos · Answer 6 · 2016-01-10T20:02:08.253

52

I had the same issue and this piece of code worked out for me.

names(data)[names(data) == "oldVariableName"] <- "newVariableName"

In short, this code does the following:

names(data) looks into all the names in the dataframe (data)

[names(data) == oldVariableName] extracts the variable name (oldVariableName) you want to get renamed and <- "newVariableName" assigns the new variable name.

edited Jan 10 '16 at 20:02

answered Jan 10 '16 at 19:37

Desta Haileselassie Hagos

23,140
7
48
53

how would this work if you had a vector with e.g. 3 oldVariableNames? – jiggunjer Dec 07 '18 at 12:14
Exactly what I was looking for -> 2 thumbs up !! – SilSur Jun 05 '20 at 16:07
Nice, not index hard-coded. – Jia Gao May 25 '22 at 00:04

score 28 · Answer 7 · edited Jan 29 '15 at 23:49

28

Similar to the others:

cols <- c("premium","change","newprice")
colnames(dataframe) <- cols

Quite simple and easy to modify.

edited Jan 29 '15 at 23:49

John Saunders

160,644
26
247
397

answered Oct 31 '14 at 22:52

Adam Erickson

6,027
2
46
33

score 23 · Answer 8 · answered Feb 21 '19 at 08:38

23

Use this to change column name by colname function.

colnames(newprice)[1] = "premium"
colnames(newprice)[2] = "change"
colnames(newprice)[3] = "newprice"

answered Feb 21 '19 at 08:38

Sophanna

347
2
7

score 13 · Answer 9 · answered Jan 25 '16 at 13:12

If you need to rename not all but multiple column at once when you only know the old column names you can use colnames function and %in% operator. Example:

df = data.frame(bad=1:3, worse=rnorm(3), worst=LETTERS[1:3])

   bad      worse    worst
1   1 -0.77915455       A
2   2  0.06717385       B
3   3 -0.02827242       C

Now you want to change "bad" and "worst" to "good" and "best". You can use

colnames(df)[which(colnames(df) %in% c("bad","worst") )] <- c("good","best")

This results in

  good      worse  best
1    1 -0.6010363    A
2    2  0.7336155    B
3    3  0.9435469    C

That code assumes the ordering of your column names equal the ordering of the inserts — Hillary Sanders, Mar 03 '16 at 19:04

score 13 · Answer 10 · answered Mar 13 '18 at 09:15

13

My column names is as below

colnames(t)
[1] "Class"    "Sex"      "Age"      "Survived" "Freq"

I want to change column name of Class and Sex

colnames(t)=c("STD","Gender","AGE","SURVIVED","FREQ")

answered Mar 13 '18 at 09:15

Mehul Katara

907
11
32

sbha · Answer 11 · 2018-11-03T12:05:28.650

There are a couple options with dplyr::rename() and dplyr::select():

library(dplyr)

mtcars %>% 
  tibble::rownames_to_column('car_model') %>%                            # convert rowname to a column. tibble must be installed.
  select(car_model, est_mpg = mpg, horse_power = hp, everything()) %>%   # rename specific columns and reorder
  rename(weight = wt, cylinders = cyl) %>%                               # another option for renaming specific columns that keeps everything by default
  head(2)
      car_model est_mpg horse_power cylinders disp drat weight  qsec vs am gear carb
1     Mazda RX4      21         110         6  160  3.9  2.620 16.46  0  1    4    4
2 Mazda RX4 Wag      21         110         6  160  3.9  2.875 17.02  0  1    4    4

There are also three scoped variants of dplyr::rename(): dplyr::rename_all() for all column names, dplyr::rename_if() for conditionally targeting column names, and dplyr::rename_at() for select named columns. The following example replaces spaces and periods with an underscore and converts everything to lower case:

iris %>%  
  rename_all(~gsub("\\s+|\\.", "_", .)) %>% 
  rename_all(tolower) %>% 
  head(2)
  sepal_length sepal_width petal_length petal_width species
1          5.1         3.5          1.4         0.2  setosa
2          4.9         3.0          1.4         0.2  setosa

dplyr::select_all() can also be used in a similar way:

iris %>%  
  select_all(~gsub("\\s+|\\.", "_", .)) %>% 
  select_all(tolower) %>% 
  head(2)
  sepal_length sepal_width petal_length petal_width species
1          5.1         3.5          1.4         0.2  setosa
2          4.9         3.0          1.4         0.2  setosa

See dplyr::rename_with() for updated tidyverse approaches as well. — Bryan Shalloway, Jul 29 '21 at 17:29

score 10 · Answer 12 · edited Jul 01 '13 at 15:12

10

try:

names(newprice) <- c("premium", "change", "newprice")

edited Jul 01 '13 at 15:12

Smern

18,746
21
72
90

answered Jul 01 '13 at 14:52

ngamita

329
2
12

score 10 · Answer 13 · answered Dec 10 '15 at 13:01

Just to correct and slightly extend Scott Wilson answer.
You can use data.table's setnames function on data.frames too.

Do not expect speed up of the operation but you can expect the setnames to be more efficient for memory consumption as it updates column names by reference. This can be tracked with address function, see below.

library(data.table)
set.seed(123)
n = 1e8

df = data.frame(bad=sample(1:3, n, TRUE), worse=rnorm(n))
address(df)
#[1] "0x208f9f00"
colnames(df) <- c("good", "better")
address(df)
#[1] "0x208fa1d8"
rm(df)

dt = data.table(bad=sample(1:3, n, TRUE), worse=rnorm(n))
address(dt)
#[1] "0x535c830"
setnames(dt, c("good", "better"))
address(dt)
#[1] "0x535c830"
rm(dt)

So if you are hitting your memory limits you may consider to use this one instead.

score 8 · Answer 14 · answered Jan 31 '13 at 19:59

8

You can just do the editing by:

newprice <- edit(newprice)

and change the column name manually.

answered Jan 31 '13 at 19:59

Baykal

569
2
10
15

Doesn't this work only for vector and factor elements? `> locanatmodelset<-edit(locanatmodelset) Error in edit.data.frame(locanatmodelset) : can only handle vector and factor elements` – vagabond Jul 27 '14 at 17:04
It works for data frames at least. That's what I know. – Baykal Jul 30 '14 at 01:25

score 3 · Answer 15 · edited Feb 13 '19 at 21:24

3

This may be helpful:

rename.columns=function(df,changelist){
  #renames columns of a dataframe
  for(i in 1:length(names(df))){
    if(length(changelist[[names(df)[i]]])>0){
      names(df)[i]= changelist[[names(df)[i]]]
    }
  }
  df
}

# Specify new dataframe
df=rename.columns(df,list(old.column='new.column.name'))

edited Feb 13 '19 at 21:24

Dave Gruenewald

5,329
1
23
35

answered May 19 '16 at 23:02

Chris

1,219
2
11
21

score 2 · Answer 16 · answered Jan 03 '18 at 08:11

2

In case we have 2 dataframes the following works

 DF1<-data.frame('a', 'b')
 DF2<-data.frame('c','d')

We change names of DF1 as follows

 colnames(DF1)<- colnames(DF2)

answered Jan 03 '18 at 08:11

Raghavan vmvs

1,213
1
10
29

score 2 · Answer 17 · edited Dec 03 '21 at 22:56

2

Change data frame column name

colnames(dataset)[colnames(dataset) == 'name'] <- 'newcolumnname'

edited Dec 03 '21 at 22:56

Henry Ecker

34,399
18
41
57

answered Nov 02 '21 at 19:20

Rupesh Kumar

157
3

score 1 · Answer 18 · answered Jan 03 '22 at 19:32

1

One option using data.table:

library(data.table)

setDT(dataframe)

setnames(dataframe,'Old1','New1')
setnames(dataframe,'Old2','New2')

answered Jan 03 '22 at 19:32

kakarot

376
4
13

score -1 · Answer 19 · answered Mar 26 '23 at 21:11

-1

This works for me:

newprice<- setNames(newprice, c("premium", "change", "newprice"))

answered Mar 26 '23 at 21:11

Abdelah

9
1

Changing column names of a data frame

19 Answers19

Linked

Related