@benedikt-buchert thanks for the update, that's a good tip. Bad graphs
most often happen due to incorrect data or issues with data ingestion.
There is some detective work involved to find the Dataset(s) causing the
merges.The following may help:Query one of the known IDs that are
causing issues using Identity Graph ViewerGrouping ID in relation to
other IDs by count. This could be done on raw data pre-ingestion or in
Query Service. This usually uncovers the issue.
From a pure tech point of view, tt can be achieved if you collect the
time and day as part of the data collection.e.g. if you are using Adobe
Analytics server side forwarding then a plugin or client-side code can
be used to populate time or day variables here The use-case is important
if you want to retarget users who have visited on a Monday then creating
the Traits and Segments in Audience Manager using the above is
appropriate. i.e. The users will qualify from the first Monday until the
Web SDK is typically used as an unauthenticated data collection source.
The data in motion would be secure viaHTTPS/TLS. Depending on your
use-case, you can also look at Edge API or streaming API as an
authenticated server to server endpoint to stream the web data.
It will depend if the data collection is for AEP or Adobe Analytics. For
AEP as compared with Adobe Analytics, where events would be assigned
based on event1,event2,etcYou would have to create a schema to capture
the event as a string value or a measure data type where the value is 1
or 0.A good starting point is to review the out of the box Field Groups
otherwise you can always create your Schemas from scratch.