It works!

I can do these patterns:

MySchema:(link,key,name)_select * from other;

And that script will select the columns named link,key and name from the other table. The schema will re-run for every new row. The result is placed correctly in nested order. I use the same nested calls for table writes as I do for command line. Tested and verified.

So, there, you have it, simple graph traversal data base. Still a lab novelty. I will try to get updates to the engine page within a day.