Skip to contents

Drop superseded and duplicate 990-PF filings, keeping every grant reported

Source: R/efilepf-dedup.R
pf_efilepf_dedup.Rd

Deduplicates at the level of the filing, never the grant row, and only within one reporting context — see the file header for why this differs deliberately from panel990::panel_deduplicate().

Usage

pf_efilepf_dedup(d, content_cols = pf_grant_content_cols(), verbose = TRUE)

Arguments

d

A grants-paid frame carrying the ef2 filing metadata: EIN2, TAX_YEAR, OBJECTID, and where available RETURN_GROUP_X, RETURN_PARTIAL_X, RETURN_AMENDED_X, RETURN_TIME_STAMP.

content_cols

Grant fields used to decide whether two filings report the same thing. Defaults to pf_grant_content_cols().

verbose

Print the tally.

Value

d with superseded and duplicate filings removed, carrying two new columns: dedup_action (kept, kept_amended, kept_flagged) and dedup_flag — TRUE where several filings of one context disagree and no amendment settles it, so all of them were kept. An attribute "pf_dedup" holds the per-context audit table.