We manually updated a bunch of subaddress records in late 2021 to add hyphens to Queens-style addresses. These appear to be incorrect but are being added by the code in this repo.
In the source geocoded data these records have values like:
F1_GRC: 1
F1_REASONCODE: 2
F1_MESSAGE: ADDR NUMBER ALTERED: HYPHEN INSERTED
HNUM: 4709
CORRECTED_HOUSE_NUMBER:
I interpret this to mean that in order to geocode the input 3rd party data the geocoder transparently inserted a hyphen into the house number. But that correction is nowhere in the geocoded delivery that we then work with in this repository. The house number should be 47-09 but we are adding 4709.
For future deliveries check if this combination of values is still present. If yes, check for a corrected house number. If there is no corrected house number I think our best plan is to filter these records out of the geocoded delivery to avoid adding more bad subaddress records.
We manually updated a bunch of subaddress records in late 2021 to add hyphens to Queens-style addresses. These appear to be incorrect but are being added by the code in this repo.
In the source geocoded data these records have values like:
I interpret this to mean that in order to geocode the input 3rd party data the geocoder transparently inserted a hyphen into the house number. But that correction is nowhere in the geocoded delivery that we then work with in this repository. The house number should be 47-09 but we are adding 4709.
For future deliveries check if this combination of values is still present. If yes, check for a corrected house number. If there is no corrected house number I think our best plan is to filter these records out of the geocoded delivery to avoid adding more bad subaddress records.