dplyr::select() with some variables that may not exist in the data frame?

Question

I have a helper function (say foo()) that will be run on various data frames that may or may not contain specified variables. Suppose I have

library(dplyr)
d1 <- data_frame(taxon=1,model=2,z=3)
d2 <- data_frame(taxon=2,pss=4,z=3)

The variables I want to select are

vars <- intersect(names(data),c("taxon","model","z"))

that is, I'd like foo(d1) to return the taxon, model, and z columns, while foo(d2) returns just taxon and z.

If foo contains select(data,c(taxon,model,z)) then foo(d2) fails (because d2 doesn't contain model). If I use select(data,-pss) then foo(d1) fails similarly.

I know how to do this if I retreat from the tidyverse (just return data[vars]), but I'm wondering if there's a handy way to do this either (1) with a select() helper of some sort (tidyselect::select_helpers) or (2) with tidyeval (which I still haven't found time to get my head around!)

mt1022 · Accepted Answer · 2021-11-19T08:26:52.923

26

Another option is select_if:

d2 %>% select_if(names(.) %in% c('taxon', 'model', 'z'))

# # A tibble: 1 x 2
#   taxon     z
#   <dbl> <dbl>
# 1     2     3

select_if is superseded. Use any_of instead:

d2 %>% select(any_of(c('taxon', 'model', 'z')))
# # A tibble: 1 x 2
#   taxon     z
#   <dbl> <dbl>
# 1     2     3

type ?dplyr::select in R and you will find this:

These helpers select variables from a character vector:

all_of(): Matches variable names in a character vector. All names must be present, otherwise an out-of-bounds error is thrown.

any_of(): Same as all_of(), except that no error is thrown for names that don't exist.

edited Nov 19 '21 at 08:26

answered Jul 26 '18 at 01:04

mt1022

16,834
5
48
71

1

all answers are good, but this one is compact and readable and doesn't throw any warnings. (@G.Grothendieck's is closest to the method I gave). – Ben Bolker Jul 26 '18 at 17:56
1

or: `d2 %>% select(which(names(.) %in% c('taxon', 'model', 'z')))` – tjebo Jul 26 '18 at 21:00
2

always good to have alternatives, but `select_if(...)` seems better than `select(which(...))` – Ben Bolker Jul 28 '18 at 21:49
`select_if (names(.)...` is very, very slow for larger data.frames. – Inferrator Dec 08 '20 at 23:49

score 7 · Answer 2 · answered Jul 26 '18 at 00:18

7

You can use one_of(), which gives a warning when the column is absent but otherwise selects the correct columns:

d1 %>%
    select(one_of(c("taxon", "model", "z")))
d2 %>%
    select(one_of(c("taxon", "model", "z")))

answered Jul 26 '18 at 00:18

Marius

58,213
16
107
105

`one_of()` is retired in favor of `any_of()` (applicable here) or `all_of()` (which would error if a variable is not present). – hplieninger Mar 26 '20 at 11:28

score 6 · Answer 3 · answered Jul 26 '18 at 00:44

Using the builtin anscombe data frame for the example noting that z is not a column in anscombe :

anscombe %>% select(intersect(names(.), c("x1", "y1", "z")))

giving:

   x1    y1
1  10  8.04
2   8  6.95
3  13  7.58
4   9  8.81
5  11  8.33
6  14  9.96
7   6  7.24
8   4  4.26
9  12 10.84
10  7  4.82
11  5  5.68

dplyr::select() with some variables that may not exist in the data frame?

3 Answers3

Linked

Related