Converting two columns of a data frame to a named vector

Question

I need to convert a multi-row two-column data.frame to a named character vector. My data.frame would be something like:

dd = data.frame(crit = c("a","b","c","d"), 
                name = c("Alpha", "Beta", "Caesar", "Doris")
                )

and what I actually need would be:

whatiwant = c("a" = "Alpha",
              "b" = "Beta",
              "c" = "Caesar",
              "d" = "Doris")

score 91 · Accepted Answer · answered Oct 09 '13 at 07:12

Use the names function:

whatyouwant <- as.character(dd$name)
names(whatyouwant) <- dd$crit

as.character is necessary, because data.frame and read.table turn characters into factors with default settings.

If you want a one-liner:

whatyouwant <- setNames(as.character(dd$name), dd$crit)

score 73 · Answer 2 · edited Sep 24 '19 at 18:34

73

You can also use deframe(x) from the tibble package for this.

tibble::deframe()

It converts the first column to names and second column to values.

edited Sep 24 '19 at 18:34

merv

67,214
13
180
245

answered Jun 06 '19 at 14:20

John Waller

2,257
4
21
26

8

Thank you John! I am amazed that after years of using the tidyverse I still learn about functions that are so useful. I wish I knew about this a long time ago. I guess I'm always a student. – HowYaDoing Oct 08 '19 at 13:47
7

For anyone coming to this post asking the inverse question - *How do I convert a named vector to a two-column data frame?* - the answer is `tibble::enframe()` – acvill Dec 10 '21 at 16:28

stevec · Answer 3 · 2021-01-26T11:48:07.100

26

Here is a very general, easy, tidy way:

library(dplyr)

iris %>%
  pull(Sepal.Length, Species)

The first argument is the values, the second argument is the names.

edited Jan 26 '21 at 11:48

answered Apr 28 '20 at 11:00

stevec

41,291
27
223
311

2

Very neat. Probably will become my goto once it goes live. – merv May 19 '20 at 03:22
2

should be the top answer in 2021 – ChrKoenig Nov 09 '21 at 19:48

alexwhan · Answer 4 · 2013-10-09T07:35:30.153

25

You can make a vector from dd$name, and add names using names(), but you can do it all in one step with structure():

whatiwant <- structure(as.character(dd$name), names = as.character(dd$crit))

edited Oct 09 '13 at 07:35

answered Oct 09 '13 at 07:13

alexwhan

15,636
5
52
66

score 8 · Answer 5 · answered Oct 09 '13 at 07:15

8

For variety, try split and unlist:

unlist(split(as.character(dd$name), dd$crit))
#        a        b        c        d 
#  "Alpha"   "Beta" "Caesar"  "Doris"

answered Oct 09 '13 at 07:15

A5C1D2H2I1M1N2O1R2T1

190,393
28
405
485

1

It should be pointed out that this makes duplicated names unique by appending a number to them. It's also not very efficient with big vectors. – Roland Oct 09 '13 at 08:14

score 3 · Answer 6 · answered May 19 '20 at 03:05

There's also a magrittr solution to this via the exposition pipe (%$%):

library(magrittr)

dd %$% set_names(as.character(name), crit)

Minor advantage over tibble::deframe is that one doesn't have to have exactly a two-column frame/tibble as argument (i.e., avoid a select(value_col, name_col) %>%).

Note that the magrittr::set_names versus base::setNames is exchangeable. I simply prefer the former since it matches "set_(col|row)?names".

Converting two columns of a data frame to a named vector

6 Answers6

Linked

Related