I'm working on a dynamic implementation of ud.convert to convert a large dataset with several different units to centimeters, and I ran into an issue. See reproducible example below where everything is converted to inches even when the conversion should be from 'centimeters' to 'cm.'
library(dplyr)
library(udunits2)
fake_data <- tibble::tibble(
trial_name = c("Trial_A", "Trial_B", "Trial_C", "Trial_D", "Trial_E", "Trial_F"),
diameter = c(2, 5, 3, 7.5, 10, 4.2),
diameter_units = c("inches", "centimeters", "inches", "centimeters", "inches", "centimeters")
)
fake_data %>%
mutate(datcm = ud.convert(diameter,diameter_units,'cm'))
Output:
trial_name diameter diameter_units datcm
<chr> <dbl> <chr> <dbl>
1 Trial_A 2 inches 5.08
2 Trial_B 5 centimeters 12.7
3 Trial_C 3 inches 7.62
4 Trial_D 7.5 centimeters 19.0
5 Trial_E 10 inches 25.4
6 Trial_F 4.2 centimeters 10.7
I can of course fix with explicit ifelse but I wanted this to work more generally. Any ideas on why this is happening would be appreciated!
A comment responding to a similar question points out that ud.convert is not vectorized over the target units argument (u2). You can create a wrapper with Vectorize().
library(dplyr)
library(udunits2)
ud.convert.vector <- Vectorize(ud.convert)
fake_data |>
mutate(datcm = ud.convert.vector(diameter, diameter_units, 'cm'))
#> # A tibble: 6 × 4
#> trial_name diameter diameter_units datcm
#> <chr> <dbl> <chr> <dbl>
#> 1 Trial_A 2 inches 5.08
#> 2 Trial_B 5 centimeters 5
#> 3 Trial_C 3 inches 7.62
#> 4 Trial_D 7.5 centimeters 7.5
#> 5 Trial_E 10 inches 25.4
#> 6 Trial_F 4.2 centimeters 4.2