new variable doesnt read well NA

[juandmaz]

I have these 2 variables, I need to modify the first one to recode correctly the numbers that are really dates and then calculate the distance in days between both columns and finally create a new variable where it says if each observation returns in the indicated time or not (975,75 days).

# A tibble: 15 x 2
   time1     time2      
   <chr>    <dttm>             
 1 44358    2015-07-23 00:00:00
 2 NO       2015-10-25 00:00:00
 3 NO       2015-05-28 00:00:00
 4 44328    2015-07-27 00:00:00
 5 NO       2015-07-24 00:00:00
 6 44277    2015-07-27 00:00:00
 7 NO       2015-07-23 00:00:00
 8 NO       2015-08-03 00:00:00
 9 NO       2015-08-03 00:00:00
10 NO       2015-07-27 00:00:00
11 NO       2015-08-03 00:00:00
12 44481    2015-07-24 00:00:00
13 NO       2015-07-24 00:00:00
14 43326    2015-07-23 00:00:00
15 44414    2015-08-03 00:00:00

I do this:

df %>%
  mutate(time1= if_else(str_detect(time1, "NO"),
                           time1,
                           time1 %>%
                             as.numeric() %>%
                             excel_numeric_to_date() %>%
                             as.character())) %>%
  mutate(dias= as.numeric(difftime(as.Date(time1), time2, units = "days"))) %>%
  mutate(comeback= if_else(any(dias>((365.25*3)-120)),T,F))

but for some reason in the created variable the NA observations of time1 are passed as T, when they should be passed as NA.

# A tibble: 15 x 4
   time1       time2               dias   comeback
   <chr>      <dttm>              <dbl> <lgl>       
 1 2021-06-11 2015-07-23 00:00:00  2150 TRUE        
 2 NO         2015-10-25 00:00:00    NA TRUE        
 3 NO         2015-05-28 00:00:00    NA TRUE        
 4 2021-05-12 2015-07-27 00:00:00  2116 TRUE        
 5 NO         2015-07-24 00:00:00    NA TRUE        
 6 2021-03-22 2015-07-27 00:00:00  2065 TRUE        
 7 NO         2015-07-23 00:00:00    NA TRUE        
 8 NO         2015-08-03 00:00:00    NA TRUE        
 9 NO         2015-08-03 00:00:00    NA TRUE        
10 NO         2015-07-27 00:00:00    NA TRUE        
11 NO         2015-08-03 00:00:00    NA TRUE        
12 2021-10-12 2015-07-24 00:00:00  2272 TRUE        
13 NO         2015-07-24 00:00:00    NA TRUE        
14 2018-08-14 2015-07-23 00:00:00  1118 TRUE        
15 2021-08-06 2015-08-03 00:00:00  2195 TRUE        
Warning message:
Problem while computing `hpv_post = if_else(...)`.
i NAs introduced by coercion

What could be happening?

[FJCC]

I think the problem is caused by the any() function. Try this version

df %>%
  mutate(time1= if_else(str_detect(time1, "NO"),
                        time1,
                        time1 %>%
                          as.numeric() %>%
                          excel_numeric_to_date() %>%
                          as.character())) %>%
  mutate(dias= as.numeric(difftime(as.Date(time1), time2, units = "days"))) %>%
  mutate(comeback= if_else(dias > (365.25*3)-120,T,F))