- Create the processing to be called as a separate model,
- Organise the data to be passed onto branches of the graph,
- Use a processing node:
- Use the operation, Insert from Model, specifying the node name, data element, justification should include elements of data from root to that node name (important), set Pass all data to No,
- Within the processing node set the property Single Match to No.
- How this works:
- The matching engine will match each branch (because multiple match allowed) and independently call the second model with the data in the match.
- Once completed the results will be combined into one resultant knowledge graph within the processing node of the calling model.
- At present this cannot be done directly, however the same result can be achieved
- Create the 2nd part as a another model and call from both points
- Call multiple times with sub-selected
- Call another model, use selection as the input, and don't select the “use all data”
- Each match will invoke the specified model with just that data… thus splitting the data and calling multiple times - similar to a “for each…”