Skip to contents

Normalize a query or reference frame

Source: R/normalize.R
np_normalize.Rd

Brings entity names and addresses onto a common, matchable form (parity with the BMF org_name_join field) and extracts the token-level features the veto layer needs. Adds columns alongside the canonical ones:

Usage

np_normalize(data)

Arguments

data

An np_query or np_reference frame from npmatch::np_query() / npmatch::np_reference().

Value

The same frame with the normalization columns added.

Details

  • name_key cleaned, abbreviation-standardized, legal-suffix-stripped

  • dba_key the DBA name under the same normalization (for cross-matching)

  • name_full cleaned name with suffix retained

  • name_core exact-lookup key: name_full without connector words and legal forms anywhere (THE, OF, AND, INC, LLC ...), keeping FOUNDATION, TRUST, FUND (see np_name_core())

  • name_placeholder TRUE for names that are not names ("SEE SCHEDULE O", "N/A", no letters); npmatch::np_block() gives such query rows no candidates

  • name_form detected legal form (INC / LLC / CORP / ...), or NA

  • name_gen, name_gen_rank vestigial — a person-name generation marker (JR / SR, ranked SR = 1, JR = 2) inherited from npmatch's person-matching lineage. Organisations do not have generations; nothing in the package consumes these and they are not useful features. Retained only for column compatibility with existing labeled extracts. See npmatch::np_default_rules().

  • name_nums space-joined embedded digit tokens (chapter/local numbers)

  • name_ord canonical ordinal rank(s) found in the name

  • street_key USPS-standardized street with unit stripped

  • street_num leading house / box number

  • street_name street body with the number, type, and unit removed

  • street_type trailing street type (ST / AVE / BLVD / WAY / ...), or NA

  • street_unit the suite / floor / unit as written, normalized to "" ("STE 300", "FL 2", "# 12"). Read from the street line, or from the unit column when the street line names none

  • street_unit_id the identifier alone, separators dropped ("STE D-2" -> "D2"). Two records are compared on this, so a BMF "STE 300" and a source "# 300" are recognised as one office

  • is_po_box TRUE if the street is a mail receptacle with an identifier, else FALSE. Any spelling of the designator (PO BOX 12, P O BOX 12, POBOX 12, POB 12, PO B 12, PO 12, BOX 12, PO BX 12, PB 12, PB BOX 12, P B0X 12, POSTAL OFC BOX 12, APDO 12, DRAWER B, PO DRAWER D, POSTAL DRAWER 12) and any identifier – digits, or a single letter where the key ends, so the lettered boxes PO BOX A through PO BOX Z count as boxes too. One of the two gates that blank the address key in np_addr_key()

  • is_mail_only TRUE if the street names a mail-delivery point that has no building behind it and no identifier this package will reduce to a box: GENERAL DELIVERY, GEN DEL, military mail (PSC 303 BOX 24, CMR 467 BOX 7000, APO, FPO, DPO) and a university mail drawer named by two letters (DRAWER JJ EMORY UNIVERSITY). The other gate on np_addr_key(), and disjoint from is_po_box by construction, so the two flags partition the buildingless addresses. Unlike a box the key is not rewritten, because there is no box number to rewrite it to. A rural route is deliberately not here: it is a unique delivery point, so RR 1 BOX 8 keeps its key, and measured that is safe – the most crowded route-with-box address in the BMF holds 5 EINs

  • mail_key the box plus its ZIP ("PO BOX 1234 59801"), or "" where the record names no identified box. The complement of the address key: blanking a box says two filers are not in one building, which is true and is all it says, while two filers collecting mail from the same box are likely administered by the same people. Weight it by np_mail_freq() before believing it – one BMF box holds 3,626 EINs. See np_mail_key()

  • street_has_box TRUE if the street carries a box component, whether or not the whole street is one, so is_po_box implies street_has_box. 211 EMMET ST S PO BOX, 300 KENAN CENTER DR CAMPUS BOX and ONE HARRISON PLZ BOX 5079 are all TRUE while is_po_box is FALSE on each. Filers write the street on line 1 and carry the box number alone on line 2; this is the column that says a bare number from that line may be a box rather than a door. It does not claim that it is – the street can name a suite too (4347 W NORTHWEST HWY STE 130 BOX), and that call belongs to whoever holds the second line. street_unit is not changed by it

  • zip5 five-digit ZIP, leading zeros restored (nested prefix)

  • zip3 three-digit ZIP prefix / SCF region (nested prefix)

  • zip9 full xxxxx-xxxx ZIP when the +4 is known, else NA

The ZIP match hierarchy is the nested prefix family zip3 \(\subset\) zip5 \(\subset\) zip9. The +4 add-on (zip_plus4) is an input only, consumed to build zip9; it is not a standalone level.

  • state_abb two-letter state code (mapped from a full state name if given)