(Geo)Data sets do share some common properties like the topic or groups of users which are authorized to read it. Often a user is interested in exactly one data set and all its Things/Locations/Datastreams, e.g. to display on a map. Currently the SensorThings API does not have a concept of "data sets" as a group or collection of all Things, Locations, Datastreams, FOIs, ... which are associated. Is it possible to implement this as an optional part of FROST-Servers functionality?
The "data set" could be part of the URL-path: {data set}/v1.1/Datastreams(x). Maybe the {data set} path part could be based on a specific properties field in entity sets. Another way could be to use PostgreSQLs schemas as namespaces for datasets.
Such functionality could be helpful in some ways:
It would make writing queries easier for users if they are only interested in a specific dataset. Currently our users need to use $filter=properties/serviceName on Datastreams to filter for entity collections of a specific dataset.
When implemented in such a way it would be easier to subscribe to topics which are part of a dataset, like {data set}/Datastreams(x)/Observations.
When using the data set name as part of the URL path it could be possible to implement access rights per dataset with one FROST-Server instance. Currently for each data set with differing access rights we would deploy a new FROST-Server instance with its own domain.
(Geo)Data sets do share some common properties like the topic or groups of users which are authorized to read it. Often a user is interested in exactly one data set and all its Things/Locations/Datastreams, e.g. to display on a map. Currently the SensorThings API does not have a concept of "data sets" as a group or collection of all Things, Locations, Datastreams, FOIs, ... which are associated. Is it possible to implement this as an optional part of FROST-Servers functionality?
The "data set" could be part of the URL-path: {data set}/v1.1/Datastreams(x). Maybe the {data set} path part could be based on a specific properties field in entity sets. Another way could be to use PostgreSQLs schemas as namespaces for datasets.
Such functionality could be helpful in some ways: