How to add multiple columns to a tibble? - r

Question

I want to add multiple empty columns to a tibble.
The names of the new columns are stored in 'columnsToAdd'
> columnsToAdd
[1] "column1" "column2" "column3" "column4" "column5"
When I run the following code lines, ...
library(dplyr)
someTibble <- tibble(name = paste("Name", 1:10))
columnsToAdd <- paste("column", 1:30, sep = "")
someTibble %>%
tibble::add_column(columnsToAdd = NA)
... I get this result, ...
# A tibble: 10 x 2
name columnsToAdd
<chr> <lgl>
1 Name 1 NA
2 Name 2 NA
3 Name 3 NA
4 Name 4 NA
5 Name 5 NA
6 Name 6 NA
7 Name 7 NA
8 Name 8 NA
9 Name 9 NA
10 Name 10 NA
... instead, I want to get the following result:
# A tibble: 10 x 6
name column1 column2 column3 column4 column5
<chr> <dbl> <dbl> <dbl> <dbl> <dbl>
1 Name 1 NA NA NA NA NA
2 Name 2 NA NA NA NA NA
3 Name 3 NA NA NA NA NA
4 Name 4 NA NA NA NA NA
5 Name 5 NA NA NA NA NA
6 Name 6 NA NA NA NA NA
7 Name 7 NA NA NA NA NA
8 Name 8 NA NA NA NA NA
9 Name 9 NA NA NA NA NA
10 Name 10 NA NA NA NA NA

Answer 1

Just create them as columns. This works for base data frames and data.table tables:
Eg:
> someTibble <- tibble(name = paste("Name", 1:10))
> columnsToAdd <- paste("column", 1:3,sep="")
That gives me one column:
> head(someTibble)
# A tibble: 6 x 1
name
<chr>
1 Name 1
2 Name 2
3 Name 3
4 Name 4
5 Name 5
6 Name 6
Then I do:
> someTibble[,columnsToAdd]=NA
and magically the columns appear:
> head(someTibble)
# A tibble: 6 x 4
name column1 column2 column3
<chr> <lgl> <lgl> <lgl>
1 Name 1 NA NA NA
2 Name 2 NA NA NA
3 Name 3 NA NA NA
4 Name 4 NA NA NA
5 Name 5 NA NA NA
6 Name 6 NA NA NA
Note this is not really magic, it is standard base R behaviour since from before R was born.

Answer 2

We can pass the column names as a list then use the !!! "triple bang"
columnsToAdd <- paste("column", 1:5, sep = "")
someTibble %>%
tibble::add_column(!!!set_names(as.list(rep(NA, length(columnsToAdd))),nm=columnsToAdd))
# A tibble: 6 x 6
name column1 column2 column3 column4 column5
<chr> <lgl> <lgl> <lgl> <lgl> <lgl>
1 Name 1 NA NA NA NA NA
2 Name 2 NA NA NA NA NA
3 Name 3 NA NA NA NA NA
4 Name 4 NA NA NA NA NA
5 Name 5 NA NA NA NA NA
6 Name 6 NA NA NA NA NA
However, I think #MrGumble may miss something as from add_column {tibble}
... Name-value pairs, passed on to tibble(). All values must have one element for each row in the data frame, or be of length 1. These arguments are passed on to tibble(), and therefore also support unquote via !! and unquote-splice via !!!.
and here is an example from tibble {tibble}
tibble(!!! list(x = rlang::quo(1:10), y = quote(x * 2)))

Answer 3

What you are seeing in tibble::add_column(columnsToAdd = NA) is the quasi-something evaluation that dplyr and tidyr introduced. If you check the definition:
> args(add_column)
function (.data, ..., .before = NULL, .after = NULL)
you'll see that it doesn't expect a certain variable. It literally expects the actual variable name, without quotation.
An entirely different approach is to create a matrix (or data.frame, whatever tickles your fancy), and smack it onto the side of someTibble:
extra <- matrix(NA_real_, nrow=nrow(someTibble), ncol=length(columnsToAdd), dimnames=list(NULL, columnsToAdd))
dplyr::bind_cols(someTibble, as.data.frame(extra))