Skip to contents

Match raw column names to canonical variable names

Source: R/match_columns.R
match_columns.Rd

Thin orchestrator: passes column names through three matching layers in descending confidence order (exact → fuzzy → keyword) and returns a structured result. Each layer is implemented as a separate internal helper so individual layers can be tested or extended independently.

Usage

match_columns(raw_colnames, config, fuzzy_max_distance = 0.15)

Arguments

raw_colnames

Character vector of column names from the raw dataset (after name normalisation).

config

Config object returned by load_config().

fuzzy_max_distance

Numeric. Maximum edit distance for fuzzy matching. Default 0.15.

Value

A named list:

match_log

data.table with one row per matched column.

rename_vec

Named character vector old -> canonical for data.table::setnames().

unmatched

Column names that fell through all three layers.

Examples

yml <- system.file("extdata", "variable_map.yml", package = "cuci")
cfg <- load_config(yml)
match_columns(c("kjonn", "alder", "unknown_col"), cfg)$match_log
#>    raw_name canonical method confidence needs_review
#>      <char>    <char> <char>     <char>       <lgcl>
#> 1:    kjonn     kjonn  exact       high        FALSE
#> 2:    alder     alder  exact       high        FALSE