Check if the ... columns of the df have a one-to-one relation

Source: R/summarize_data.R

Aim to test cases like:

  • A table with person_id and birth_date: each person_id must always have the same birth_date. Calling are_paired(df, person_id, birth_date) lets you test that. birth_date, of course, can be duplicated. Even person_id could also be duplicated.

  • data with region_name and region_code

are_paired(df, ...)

Arguments

df

a data.frame

...

unquoted columns of the df to test. If empty, all columns used

Value

logical

Details

TODO: let's find a better name for this TODO: check better algorithms

Examples

are_paired(mtcars)
#> 32 combinations for all columns
#> number of unique values for each column as follows:
#>  mpg  cyl disp   hp drat   wt qsec   vs   am gear carb 
#>   25    3   27   22   22   29   30    2    2    3    6 
#> [1] FALSE
are_paired(mtcars, mpg, cyl)
#> 27 combinations for all columns
#> number of unique values for each column as follows:
#> mpg cyl 
#>  25   3 
#> [1] FALSE