Hi @Anonymous ,
Responding to your question, "Is this doable?", yes, with certain considerations and assumptions about how unique these are?
For example, if we have one data model for production process. So it will have date/machine/product wise production data. And now say, we installed an automatic building management system which can log environment data every minute and supplier of this new system also make available raw data for you which you use in creating another data model. Right?
Now, though we may or may not say, that both these are unique but at least both have something in common, say at least date of production and date of environment readings, so it is natural to say that you may try to correlate production data with environment data on same time event, if there was any product defect correlated with environment change and so on.
In this case, date field was common to both the models and this is the way you need to find common items between such models. I am decribing all this on the assumption that since you had the thought to relate two models, there is a natural relation between them although they are generated independently.
The above concept is the part of dimnesion modelling and if there are more interests you can read through Kimball's dimensional modelling!
Just for reference, I also share a simple video here to take you to a practical example:
https://www.youtube.com/watch?v=TnyRsO4NJPc
Hope above answer helps you.
