Select columns by regex matching

Question

I know there a lot of posts on this, but I couldn't find any that helped. What I'm trying to do is simple.

I want to select (or drop) columns based on whether a letter is present in a column name.

library(tibble)
library(stringr)
library(dplyr)

xy <- tibble("x" = 1:5, "123" = 6:10, "x123" = 11:15)

#does not have expected output
> xy %>% select(-matches("[:alpha:]"))
# A tibble: 5 x 3
      x `123`  x123
  <int> <int> <int>
1     1     6    11
2     2     7    12
3     3     8    13
4     4     9    14
5     5    10    15

But if I use str_view_all all the results are as expected, but doesn't work with the dplyr helper function matches for selecting columns.

str_view_all(x, "[:alpha:]")

I'm looking for a solution using stringr and dplyr if possible. It seems like this should be very straightforward.

acylam · Accepted Answer · 2018-08-24 19:35:58Z

1

[:alpha:] is a POSIX character class that is only valid inside a bracketed expression. So you need an extra pair of brackets:

xy <- tibble("x" = 1:5, "123" = 6:10, "x123" = 11:15)

xy %>% select(-matches("[[:alpha:]]"))

Result:

# A tibble: 5 x 1
  `123`
  <int>
1     6
2     7
3     8
4     9
5    10

edited Aug 24, 2018 at 19:35

answered Aug 24, 2018 at 19:21

acylam

18.8k5 gold badges39 silver badges47 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Lloyd Christmas Over a year ago

facepalm. okay that works brilliantly. Just curious why you don't need the extra bracket when using str_view_all for example.

acylam Over a year ago

@LloydChristmas Good question that I do not have an answer to. [:alpha:] is what is called a POSIX character class and will only be evaluated as such inside a bracketed expression [...]. [:alpha:] by itself would just mean that you're matching the characters :, a, l, p, h literally. This is why it puzzles me why your str_view_all example works in the first place.

Collectives™ on Stack Overflow

Select columns by regex matching

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related