Check API pagination

Did you obtain every expected page?

A 200 response may contain only the first page. Pagination can use a page number, offset, cursor or next link. This guide provides a reproducible collection check; the exact mechanism comes from the service documentation.

1. Fix scope

   Keep query, filters, version and sort order. Record the announced total and whether it counts records, estimated results or distinct objects. Compare equivalent filters.

2. Follow documented continuation

   Read the supplied next link or cursor. Log every transition and detect previously visited continuations. Do not invent addresses when the service supplies its own.

3. Reconcile counts

   Count received records and unique identifiers separately. Overlapping pages inflate record totals. Retain repeated objects to inspect their versions before deduplication.

4. Document the stop

   Record declared end, collection cap, error or interruption. For changing data, use a snapshot when available or disclose the limitation. Resume with unchanged filters.

Situations and decisions

Typical situations for preparing a check. They do not describe completed assignments or actual observations.

Record to retain

• Query, filters and version
• Page transitions and responses
• Raw and unique counts
• Stop reason and resume conditions

Official references

• GitHub — REST pagination
• IETF — Web Linking, RFC 8288

Related source profiles

• data.gouv.fr
• Eurostat · Data Browser
• Crossref Metadata Search
• Companies House

Related protocols

• Query a public API and retain a verifiable extract
• Keep a reproducible version of public data
• Check a join before combining two files

Checks before reaching a conclusion

• Are filters stable?
• Are continuations unique?
• Are repetitions explained?
• Is the stopping reason explicit?

A collection log separating technical success, observed coverage and interrupted collection.