Create header of a dataframe from the first row in the data frame

Question

Suppose I have a dataframe:

a <- data.frame(a=c("f", 2, 3), b=c("g", 3, 7), c=c("h", 2, 4))

and I would like to create column names from the first row. MY guess was:

names(a) <- a[1,]

which gives:

names(a)
[1] "3" "3" "3"

I did not fully get what is happening. Can anyone explain and help me on how to do it the right way?

Why do you have this data.frame? Looks like a problem that should be solved during data import/creation. — Roland, Mar 28 '14 at 14:06
@Roland I import a csv containing factors and numerical data. therefore I can't use a matrix...if this is what you mean. — joaoal, Mar 28 '14 at 17:37
If you import a CSV, `read.csv` should recognize the column headers. — Roland, Mar 28 '14 at 20:33
yes. but the first line does not contain the actual header but metadata-blabla. Bad secondary data I have to clean up. — joaoal, Apr 03 '14 at 19:38

score 2 · Accepted Answer · answered Mar 28 '14 at 14:07

2

The columns of a are factors with each column having different levels. R casts them to ints. Since, for each column the letter appears last alphanumerically it gets assigned the value 3.

Try

names(a) = as.character(unlist(a[1,]))
names(a)

answered Mar 28 '14 at 14:07

user3472874

151
5

What is the `as.character` doing in this answer? – A5C1D2H2I1M1N2O1R2T1 Mar 28 '14 at 14:12
Nothing! Works without it entirely! – user3472874 Mar 28 '14 at 14:15
It seems that `names()` does the conversion of factor to character itself – SeDur Mar 28 '14 at 14:19
... at which point, unfortunately, this answer is the same as Jilber's, posted a couple of minutes before you, and as the answer that was posted by rawr at one point in the comments... Oh well.... – A5C1D2H2I1M1N2O1R2T1 Mar 28 '14 at 14:27
ok you saved my life...but what was happening here? I'm cleaning up a large Excel file and this example isn't helping me.. – Atieh Nov 25 '14 at 15:53

Jilber Urbina · Answer 2 · 2014-03-28T14:27:23.823

1

Try this:

> colnames(a) <- unlist(a[1,])
> a
  f g h
1 f g h
2 2 3 2
3 3 7 4

edited Mar 28 '14 at 14:27

answered Mar 28 '14 at 14:05

Jilber Urbina

58,147
10
114
138

I don't think `droplevels` is required here, is it? – A5C1D2H2I1M1N2O1R2T1 Mar 28 '14 at 14:08
You are right @AnandaMahto, It's 8:20am, it's too early to think about R, need a coffee and thanks for pointing that out. – Jilber Urbina Mar 28 '14 at 14:27
2

or 8PM for folk like me, perhaps too late to think about R :-) – A5C1D2H2I1M1N2O1R2T1 Mar 28 '14 at 14:28

score 1 · Answer 3 · answered Sep 26 '20 at 16:45

1

janitor::row_to_names(a,1)

janitor package gives the cleanest way to shift any data row up as column names.

answered Sep 26 '20 at 16:45

Lazarus Thurston

1,197
15
33

Create header of a dataframe from the first row in the data frame

3 Answers3

Linked