Adding row to a data frame with missing values

Question

So I have this data frame

df <- data.frame( A=1:10, B=LETTERS[1:10], C=letters[1:10], stringsAsFactors= F )

I want to add a row to this data frame with A=11 and C="K", but I don't have the information on the column C. In fact, I have no idea a priori what other columns are there in the data frame, except for A and B. These missing columns should get an NA. The best I could come up is this:

tmp <- rep( NA, ncol( df ) ) # Batman!
df <- rbind( df, tmp )
df[ nrow( df ), "A" ] <- 11
df[ nrow( df ), "B" ] <- "K"

Is there a simpler way?

Check out the new version(s) I added to my answer. I think that is the cleanest yet. — BrodieG, Feb 26 '14 at 18:16

BrodieG · Accepted Answer · 2014-02-26T18:11:19.320

14

Base version:

new.row <- list(A=11, B="K")
df[nrow(df) + 1, names(new.row)] <- new.row

plyr version:

library(plyr)
new.row <- data.frame(A=11, B="K", stringsAsFactors=F)
df <- rbind.fill(df, new.row)

Both produce:

    A B    C
1   1 A    a
2   2 B    b
3   3 C    c
4   4 D    d
5   5 E    e
6   6 F    f
7   7 G    g
8   8 H    h
9   9 I    i
10 10 J    j
11 11 K <NA>

You can also generalize the base version to more rows:

more.rows <- data.frame(A=15:20, B=letters[15:20], stringsAsFactors=F)
df[(nrow(df) + 1):(nrow(df) + nrow(more.rows)), names(more.rows)] <- more.rows

producing:

    A B    C
1   1 A    a
2   2 B    b
3   3 C    c
4   4 D    d
5   5 E    e
6   6 F    f
7   7 G    g
8   8 H    h
9   9 I    i
10 10 J    j
11 11 K <NA>
12 15 o <NA>
13 16 p <NA>
14 17 q <NA>
15 18 r <NA>
16 19 s <NA>
17 20 t <NA>

edited Feb 26 '14 at 18:11

answered Feb 26 '14 at 15:22

BrodieG

51,669
9
93
146

@January, I also added a `base` version. – BrodieG Feb 26 '14 at 15:24
Nice use of `match` in the base version, but note that this approach changes the integer column to character. – Matthew Plourde Feb 26 '14 at 15:30
@BrodieG as Matthew pointed out, the first example has a deficiency (there was a reason I picked up a mixture of ints and characters...) – January Feb 26 '14 at 15:43
@January, see update, though I think at this point Matthew's is cleaner. – BrodieG Feb 26 '14 at 16:48

score 6 · Answer 2 · answered Feb 26 '14 at 15:40

6

Here's an alternative base version that preserves the datatypes.

new.row <- head(df[NA,], 1)
new.row[c('A', 'B')] <- list(A=11, B='K')
rbind(df, new.row)

answered Feb 26 '14 at 15:40

Matthew Plourde

43,932
7
96
113

3

Nice; how about `df[1,][NA,]` for the first line? Then we do not generate another data frame of the same size as df. – January Feb 26 '14 at 15:46

Adding row to a data frame with missing values

2 Answers2

Linked