Skip to content

Locality parsing edge cases (was #5 in companies-house-ETL) #5

Description

@giacecco

@pezholio on 13/11

I've just noticed this in the raw DB after doing a new import

{
  "saon": "5TH FLOOR",
  "paon": "155",
  "street": "EAST BARNET ROAD",
  "locality": "NEW",
  "town": "BARNET",
  "postcode": "EN4 8QZ",
  "country": null
}

Googling the address reveals that the locality is actually New Barnet, so I guess it's assuming the 'Barnet' part of the address is a town name and parsing it out?

@MurrayData's response

The reason for this is an assumption based on Barnet being a valid postal town for the postcode and an assumption that anything between the street and the town is a locality which is true in the vast majority of cases.

However with changes agreed yesterday new validation tables and rules are to be used which will remedy this. Assumptions will be parameterised in tbe configuration file so can be tweaked.

These will be implemented next week when I'm back from Germany.

@pezholio's response

Lovely. I figured it'd be in hand, but thought I'd flag it up anyway :)

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions