PSC Resource ID

Data Issues
1

Is “resource_id” the only field which can help us identify PSCs uniquely?

Does it correspond to the “links_self” field in the bulk PSC Snapshot?

2

If you mean whether a PSC resource_id identifies the same source relationship consistently, then yes, it is intended to identify that particular PSC record.

If you mean whether it identifies the person across all Companies House records, then no. The same person may appear under multiple resource_id values, and different people may also share the same name. We therefore use corroborating evidence such as normalised name, DOB, address and a unique Officer anchor.

These figures were taken from Augusts data files:

  • 44,498,666 eligible source links before resolution
    • 30,698,283 Officer identity records
    • 13,800,383 individual PSC identity records
  • 38,845,171 Person records after resolution
  • 5,653,495 excess source-link occurrences, approximately 12.7%

RE: links_self

The official Companies House Streaming API schema shows them as separate top-level fields:

  • resource_id — “The ID of the resource”
  • resource_uri — “The URI of the resource”
  • data.links.self — “The URL of the resource”

See the official PSC Streaming API schema, especially the fields around resource_id, resource_uri, and data.links.self.

Companies House also explicitly distinguishes a PSC’s identity from a notification/relationship ID in its PSC resource specification update:

“psc_id is a unique ‘id’ for a PSC, notification_id is a unique ‘id’ for an actual ‘notification’.”

So the token at the end of links.self should not automatically be described as the same field as resource_id.

1 Like