Skip to contents

Retain only source records not already matched (incremental matching)

Source: R/fetch.R
np_diff_unmatched.Rd

Compares a fresh source table against an existing crosswalk and returns the subset of fresh records that have not already been resolved — so a re-run only spends compute on new or previously-unmatched organisations.

Usage

np_diff_unmatched(
  fresh,
  crosswalk,
  key_fresh = NULL,
  key_cross = NULL,
  status_col = NULL,
  matched_values = c("YES", "MATCH", "match"),
  quiet = FALSE
)

Arguments

fresh

Fresh source records (data frame) with an id key.

crosswalk

Existing crosswalk (data frame) with an id key and, ideally, a tier/status column.

key_fresh, key_cross

Key column names. Auto-detected from uei / UNIQUE ENTITY ID / unique_entity_id / .id if NULL.

status_col

Status/tier column in crosswalk. Auto-detected from tier / match_decision / status if NULL.

matched_values

Status values that count as already matched.

quiet

Suppress the summary message.

Value

The retained subset of fresh, with an integer "diff" attribute (fresh, already_matched, retained, new_ids).

Details

A record counts as "already matched" if its key appears in the crosswalk with a status in matched_values (when a status/tier column is present or named via status_col); if no status column is found, any appearance in the crosswalk counts as matched. Set matched_values = c("YES","MAYBE") to also retain (re-run) review-pending cases against a refreshed reference.