resolve the RAW_/ANON_-prefixed column names a score_*() call needs, and check up front that all four exist

Every score function needs exactly the same four columns, derived the same way from `target` and `row_number`. Deriving them in one place keeps the four functions (and their error messages) from drifting apart.

Usage

reid_prefixed_columns(dat_raw_anon, target, row_number, fn_name)

Arguments

dat_raw_anon

dataframe of raw_anon form

target

target column name, *before* RAW_/ANON_ prefixing

row_number

row-number column name, *before* RAW_/ANON_ prefixing

fn_name

name of the user-facing function, used in the error message so the message names the function the caller actually called

Value

a list with elements `raw_target`, `anon_target`, `raw_row_number`, `anon_row_number`