Skip to content

Knowledge Graph

All objects were stored in a graph to connect related objects together. Same objects from different source systems will be connected and at the end in the infocenter returned as a single object which is merged with the properties from all sources.


If the system is confident that two objects were similar it will automatically merge them. The object with lower confidence will be merged into the one with the better confidence / data quality. If an object is merged the old object id will be deleted. If you cache the infocenter objects or if you use the identifier e.g. to connect data you should handle these scenarios. The merging has a similar effect as if the type of an object is changing then the identifier will change too.

There are two options for how you can experience the id change:

ds-state header

If you call the GET endpoint for an identifier which was changed then the object will still be returned but the response content includes the new object with the new identifier and @id property to retrieve it from the correct endpoint. In this case, the custom header ds-state is returned with obsolete;<new object type>;<new id> (e.g. obsolete;LocalBusiness;loc_px9_ciaidd)