@gcarmich perhaps attach a sample dataset for community members to view and experiment with.

I still can't upload files to the FME Community. I sent the workflow and an image to your email. Can you post for me?

need_to_process_one_by_one.fmw

1. Create a "CircleID" Attribute on the circles with a Counter if there isn't one already
2. Create a separate "JoinField" Attribute on both the Circles and the Lines, and just set to a Dummy value like 1
3. Use FeatureJoiner to join the the Lines to the Circles to create multiple Groups of the same Lines. The output of this will be multiple sets of the same Lines, but now each set of Lines will have their own CircleID to match the individual Circle they are grouped with
4. In LineOnLineOverlayer, use the CircleID as the Group By Attribute. This ensures that each Circle and Lines Group is processed independently

not_all_points_showing.fmwI made the changes suggested. When I run it using a sample of 10 circles, the LineOnLineOverlayer takes in 30 objects (10 circles x 3 lines) but only outputs 10 intersections. Each circle should intersect each of the 3 lines once and therefore produce 3 intersection points. How do I set the LineOnLineOverlayer to output all intersections for each circle?process_one_by_one2.JPG

That's because the workspace is set to send 3 x 10 = 30 Circles to LineOnLineOverlayer. What you want it to do is send: 10 Circles + 10 x 3 Lines.

Much like this:

As an aside, you'll find AttributeCreator easier to create values with rather than StringConcantenator.

It seems to be working but it is also returning the vertex where the 3 lines come together. I can easily delete that point but I'm curious why it is appearing. I'll also need to study the process because I don't fully understand it. Thank you.

remaining center vertex.JPG

The reason is because LineOnLIneOverlayer looks for any intersections of Lines in the Group. That will include intersections between the Lines themselves. Usually these can be filtered out by these not having an Attribute that only exists on the Circles.

The methodology as to why it works is because "Group By" mode in a Transformer treats each set of Features as an independent Group of Features to be processed independently of each other. What the above workflow does is group together 3 Lines for every Circle by giving the Lines the same "CircleID"

Note, an alternate way for this particular problem that avoids Groups is to instead use a Transformer with separate Base and Candidate input ports, like NeighborFinder