General Issue: Improve `derive_interp_records()`

[bundfussr]

Background Information

• The final dataset shouldn't be assigned to a variable. Then the returned dataset is printed. At the moment we get:

  > derive_interp_records(my_cdc, by_vars = exprs(SEX), parameter = "HEIGHT")
  > 

  Then also the example in the documentation could be simplified to

  library(dplyr, warn.conflicts = FALSE)
  library(rlang)
  
  admiralpeds::cdc_htage %>%
    mutate(
      SEX = case_when(
        SEX == 1 ~ "M",
        SEX == 2 ~ "F",
        TRUE ~ NA_character_
      ),
      # Ensure first that Age unit is "DAYS"
      AGE = round(AGE * 30.4375)
    ) %>%
    # Interpolate the AGE by SEX
    derive_interp_records(
      by_vars = exprs(SEX),
      parameter = "HEIGHT"
    )

• The function drops variables which are not expected. derive_*() functions shouldn't drop any variables or records.
• Why is the uniqueness of records only checked if by_vars is NULL? I wonder if the check should be replaced by calling signal_duplicate_records(dataset, by_vars = exprs(!!!by_vars, AGE)) regardless of the value of by_vars.
• In the documentation of the dataset argument AGEU is listed as required variable. However, this isn't checked nor is the variable used in the code.
• The "Value" section of the documentation should provide more details. I had to read the code to understand what the function is doing. For example it should be stated that linear interpolation is done and that the output dataset contains a record for each day from the minimum age to the maximum age.

Definition of Done

No response

[rossfarrugia]

thanks Stefan!

@Fanny-Gautier please can you consider these comments and make updates in a new branch/PR if you agree

[Fanny-Gautier]

Unit tests are now failing since keeping potential extra variables: I have to re-work on it. Will try to complete them by tomorrow.