Str_detect multiple columns using across

Question

I would like to create a new column based on the results of str_detect across multiple columns using across.

For example, in the test data below, I'd like to search for "No job" across columns that start with "job", then return 1 if that string is detected in any of the columns, and 0 if it is not.

test_data <-  data.frame("job1" = c('Sales','Baker','Blacksmith','Brewer'), 
                         "job2" = c('Mailman','Jockey','Jobhunter',"No job"),
                         "id" = c("id_1", "id_2", "id_3", "id_4"))

# Output I'd like:

#         job1      job2   id no_job
#1      Sales   Mailman id_1      0
#2      Baker    Jockey id_2      0
#3 Blacksmith Jobhunter id_3      0
#4     Brewer    No job id_4      1

I know I could unite the columns that start with "job", and then just use str_detect on that new column like this:

test_data2 <- test_data %>%
    unite(col = "all_jobs", starts_with("job"), sep = ", ", remove = FALSE) %>%
    mutate(no_job = if_else(str_detect(all_jobs, "No job"), 1, 0))

... but I was wondering if there was a way to use across to do the same thing. I'd tried variations on the following but haven't gotten it to work.

test_data2 <- test_data %>%
    mutate(no_job = if_else(across(starts_with("job"), str_detect(., "No job")), 1, 0))

tmfmnk · Accepted Answer · 2020-08-13 16:43:34Z

10

One option could be:

test_data %>%
 rowwise() %>%
 mutate(no_job = +any(str_detect(c_across(-id), "No job")))

  job1       job2      id    no_job
  <fct>      <fct>     <fct>  <int>
1 Sales      Mailman   id_1       0
2 Baker      Jockey    id_2       0
3 Blacksmith Jobhunter id_3       0
4 Brewer     No job    id_4       1

edited Aug 13, 2020 at 16:43

answered Aug 13, 2020 at 16:38

tmfmnk

40.4k4 gold badges54 silver badges73 bronze badges

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

5 Comments

Ben Norris Over a year ago

In case "No job" appears in multiple columns, you could use mutate(no_job = as.numeric(any(str_detect(c_across(-id), "No job"))))

Dunois Over a year ago

What's the deal with the + ther in that mutate()? Could you please explain? I have never encountered that before, and I am very curious!!

tmfmnk Over a year ago

@Dunois it is the same as as.numeric(), i.e. it just converts a logical vector to a numerical one.

flxflks Over a year ago

@Dunois, is there a way to apply this to specific columns? I have a larger dataset and I would just like to apply this for specific columns.

Dunois Over a year ago

@flxflks I'm not the OP, but in general, you can use selection verbs like starts_with() inside c_across() to specify a subset of columns. Or even more generally, you could just use across() within mutate().

Dharman · Accepted Answer · 2022-03-18 23:06:40Z

2

I encountered a similar problem and this is a possible solution using case_when:

test_data %>% mutate(no_job = case_when(if_any(str_detect(starts_with("job"), "No job"))~1))

edited Mar 18, 2022 at 23:06

Dharman♦

34k27 gold badges106 silver badges158 bronze badges

answered Mar 18, 2022 at 22:58

JisL

3331 silver badge9 bronze badges

Comments

SAL · Accepted Answer · 2024-12-01 15:28:56Z

0

Another solution could be pmap from purrr family:

library(tidyverse)
test_data %>% 
mutate(no_job=pmap_int(across(1:2), ~ str_detect(c(...), "No job") %>% sum()))

output:

        job1      job2   id no_job
1      Sales   Mailman id_1      0
2      Baker    Jockey id_2      0
3 Blacksmith Jobhunter id_3      0
4     Brewer    No job id_4      1

answered Dec 1, 2024 at 15:28

SAL

2,2862 gold badges8 silver badges17 bronze badges

Collectives™ on Stack Overflow

Str_detect multiple columns using across

3 Answers 3

5 Comments

Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

5 Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related