create logical column based on multiple existing columns R

June 27, 2023

I trying to filter values over columns accordingly to some conditions.

I have the following part of my data:

df <- tibble(
    v1 = c(5, 7, 4, 2, 2, 3, 3, 4, 5, 6, 1, 7),
    v2 = c(1, 1, 1, 5, 7, 30, 1, 4, 5, 6, 6, 7),
    v3 = factor(c("a", "a", "a", "a", "a", "a", "b", "b", "b", "b", "b", "b"))
)

I trying to create a column test that verifies if each row by group (v3 in this case) are between the mean +- standard deviation (sd) of their group and return a logical value (T/F).

For example, if the first value in v1 (5) is between mean +- sd for group a in v1 and the first value in v2 (1) is between mean +- sd for group a in v2 the column test will return a TRUE in another case return FALSE.

The expected output I wait is:

# A tibble: 12 × 4
      v1    v2 v3    test 
   <dbl> <dbl> <fct> <lgl>
 1     5     1 a     TRUE 
 2     7     1 a     FALSE
 3     4     1 a     TRUE 
 4     2     5 a     TRUE 
 5     2     7 a     TRUE 
 6     3    30 a     FALSE
 7     3     1 b     FALSE
 8     4     4 b     TRUE 
 9     5     5 b     TRUE 
10     6     6 b     TRUE 
11     1     6 b     TRUE 
12     7     7 b     FALSE

I try the following code, but not return what I expected

df %>%
  mutate(teste = if_all(v1:v2, ~ . <= mean(.) - sd(.) | . >= mean(.) + sd(.), TRUE, FALSE))

Thanks for any help

>Solution :

Try this:

df %>%
  mutate(teste = if_all(v1:v2, ~ . >= mean(.) - sd(.) & . <= mean(.) + sd(.)),.by=v3)

      v1    v2 v3    teste
   <dbl> <dbl> <fct> <lgl>
 1     5     1 a     TRUE 
 2     7     1 a     FALSE
 3     4     1 a     TRUE 
 4     2     5 a     TRUE 
 5     2     7 a     TRUE 
 6     3    30 a     FALSE
 7     3     1 b     FALSE
 8     4     4 b     TRUE 
 9     5     5 b     TRUE 
10     6     6 b     TRUE 
11     1     6 b     FALSE
12     7     7 b     FALSE