It has been a while since my initial question of how to combine ecoinvent with parts of EXIOBASE. Here is what I learned so far.
When you export all processes in one country's folder as JSON-LD and tick "Export default providers," you will only get the foreground system. If you want all the processes and flows that are needed in the background, you need to create a product system that includes all processes of that country as inputs, then calculate the system, and then export this product system with "Export default providers" ticked. LCIA results are the same now for the original and the one-country database.
You can import this one-country database into the ecoinvent database. If you calculate a product system using datasets from both databases, you will by default get results for ecoinvent and EXIOBASE separately. This is because the different impact methods have different elementary flows associated. But you can fix this quite easily if you create your own impact category and insert the flows from EXIOBASE and ecoinvent. For example, you create a new impact method called "GWP100" and copy the flows from CML 2001 baseline global warming from EXIOBASE and ecoinvent into this new impact category. You can find the flows under impact categories and then the tab "Characterization factors."
Be aware that all of this takes a lot of RAM and calculation time. The PC I am using has 32 GB of RAM and I allocated openLCA 30 GB. Still, all these calculations and exports took about 12 hours.
Although there are some minor issues, especially regarding performance, I did check all the other major competitors on the market like Brightway, GaBi, Umberto, and SimaPro, and I will stick with openLCA because of cost/benefit.
Now to my question: As I said, I calculated a product system containing all processes of EXIOBASE from one country as an input. When I now analyze the metadata of my database by clicking on Database → Content → Processes / Flows, I get the following picture: Processes are about 15% less in the one-country database (8,100) compared to the original EXIOBASE (9,800). But the number of flows (77,000) is more than 7× higher than the original EXIOBASE (11,000) after the export. In the original database, none of the flows are duplicated, while in the one-country database I see a lot of duplicate elementary flows. Addtionally, there a lots of new flows that are not part of the original EXIOBASE. For example Chromium III is 57 times listed in the one-country-database but does not appear once in the original EXIOBASE. In the original database, when I click in the navigation menu on Flows → Elementary Flows, in all the subcategories like "Emissions to air," there are only flows categorized as "unspecified." After the export, there are flows in all the other folders, e.g., "Emissions to air → High population density."
Why are these new flows like e.g., Chromium III added in the export DB and where do they come from? Is it necessary to have all these duplicate elementary flows? Is there a way to easily clean a database and delete unnecessary and duplicate flows?