Skip to contents

Parse the secondary-unit line into a comparable unit

Source: R/unit.R
pf_street_unit.Rd

Extracts the unit from a raw PF_15_G_PAID_RECIP_ADDR_L2 value and returns it in npmatch's canonical street_unit spelling — "STE 400", "FL 2", "# 206" — or "" where the line holds no unit.

Usage

pf_street_unit(x, street = NULL)

Arguments

x

Character vector of raw address line 2 values.

street

Optional street line, same length as x, used for the box-number rejection. A value carrying an explicit designator is still read as a unit.

Value

A character vector the length of x, holding npmatch's canonical street_unit string or "".

Details

The parse itself is npmatch's: the value is passed to npmatch::np_normalize() as a unit column and street_unit is read back, so pfmatch compares units the same way the matcher does and a normalizer change moves both together. Values are de-duplicated before normalizing, since line 2 repeats heavily.

Two rejections are applied on top, because ADDR_L2 is not a unit field — street lines, city names, PO boxes and prose all appear in it. Both apply only to a bare identifier, one carrying no designator; a value with an explicit designator is always read as a unit.

  • No digits. STREET, WAY and CONSERVANCY are real values of this field, and npmatch's bare-identifier path reads each as a unit id. The digit requirement is pfmatch's, and it is why it belongs on the bare path only: SUITE B is a real suite with no digit in it.

  • A box number at a box street. At 211 EMMET STREET SOUTH PO BOX, the 400891 on line 2 is a box, not a door. See .PF_BOX_RE.

See also

pf_unit_evidence(), and dev/UNIT-UPSTREAM-SYNC.md for the decisions this delegation settles.