Deduplicate grant recipients — pf_dedup • pfmatch

Deduplicate grant recipients

Usage

pf_dedup(
  grants,
  method = c("exact", "fuzzy"),
  threshold = 0.92,
  max_block = 2000L,
  verbose = TRUE
)

Arguments

grants: A prepared frame from pf_prepare_grants().
method: "exact" or "fuzzy" (fuzzy implies exact first).
threshold: Jaro-Winkler similarity required to join two names in the same geographic block. Default 0.92.
max_block: Blocks larger than this are compared exactly only, not fuzzily — a guard against a quadratic blow-up on a huge shared ZIP.
verbose: Print the collapse ratio.

Value

grants with two columns added: recipient_uid (the identity a match is computed for) and recipient_n (how many grant rows share it).