federated catalog search

Federated catalog search queries many national data portals through one aggregator. The cataloged example is the EU Open Data Portal, whose short description documents federated search over data.europa.eu for 'textile' and which carries the tag 'federated-catalog'. A single 'textile' query reaches collection, recycling, trade and industry datasets held by every EU member-state portal.

The mechanism is metadata harvesting: the aggregator indexes what each national portal publishes and resolves queries against that combined index. The searcher writes one query, and the coverage comes from however many portals have been harvested.

• One query, many owners, uneven quality. Because results come from every member-state portal, completeness depends on each country's harvesting and metadata depth varies by owner.

• Coverage claims need a probe query. The documented 'textile' query is the test: if a portal's records do not appear, they are not being harvested.

• Licensing follows the source dataset. A federated search returns metadata from many publishers, so reuse terms must be read per record rather than assumed at the portal level.

1. Run the reference query. The cataloged record proves 'textile' reaches collection, recycling, trade and industry datasets across member-state portals; run your own equivalent before committing.

1. Check the harvest scope. Confirm which national portals the aggregator indexes and how recently each was harvested.

1. Read licenses per result. Distribution-level licenses differ by publisher, so a portal-level assumption will be wrong somewhere.

1. Prefer the API over the web UI. The official-api pattern is what makes federated search repeatable inside a pipeline.

These entries cover the mechanism and the access pattern behind it:

metadata harvesting is the mechanism underneath: the aggregator indexes what each member-state portal publishes.

official api is what makes a federated query repeatable inside a pipeline instead of a manual portal search.

The textile datasets reached by the reference query sit in the catalog's textiles data page.