I am trying to create a new column (age_clean) based on data in 2 columns: age (numeric) and age_unit (days, weeks, months, years).

If the age is between 0-<1 OR age unit equals weeks OR days or age is between 0-11 with unit months, have age_clean = 0.

Here is my code:

database %>%
  mutate(age_clean = 
           case_when(!age_unit %in% c("years", "days", "months", "weeks") ~ "Other",
                     (age >= 0  & age < 1) | (age_unit == "weeks" | age_unit == "days") | (age >= 0  & age <= 11 & age_unit == "months") ~ '0',
                     TRUE ~ as.numeric(age)))
    
Error in `mutate()`:
! Problem while computing `age_clean = case_when(...)`.
Caused by error in `` names(message) <- `*vtmp*` ``:
! 'names' attribute [1] must be the same length as the vector [0]
Run `rlang::last_error()` to see where the error occurred.

Note that for some of the rows, there is no data in age or age_unit. Maybe I am missing a clause for NA?

>Solution :

It looks like you overcomplicate your cases though, but I am not 100% sure if I understand you well. I assume you want "age_clean" to be an age in years? If so you can do it this way.

library(dplyr)

database %>% mutate(
  age_clean = case_when(
    age_unit == "years" ~ age,
    age_unit == "months" ~ age %/% 12,
    age_unit == "weeks" ~ age %/% 52,
    age_unit == "days" ~ age %/% 365
  )
)

Or you can use this instead of using cases at all:

mutate(age_clean = floor(as.numeric(lubridate::duration(age, units = age_unit)) / (60 * 60 * 24 * 365)))

results

#   age age_unit age_clean
# 1   2    years         2
# 2  18    years        18
# 3 100     days         0
# 4 380     days         1
# 5  10   months         0
# 6  26   months         2
# 7  25    weeks         0
# 8  54    weeks         1

data

database <- data.frame(
  age = c(2, 18, 100, 380, 10, 26, 25, 54),
  age_unit = c("years", "years", "days", "days", "months", "months", "weeks", "weeks")
)