I have a graph made of two collections, one is a collection of 60,000 services and 10 products, and another is a collection of 1 million edges connecting a product to a service. We want to find all the service vertices on our graph that are connected to all 10 products. Basically, we want to find the intersection of all ten products at a single service, and return all the services at which this occurs. Each of the ten products have a distinct product ID. Does anyone have any suggestions on how to write a somewhat efficient query that returns these intersections of products? Below are some screenshots of a small portion of each of the collections, in JSON format.
1 Answer
The following query use two collections products and edgeCol. It first collects all of the 10 products in prods, then it iterates over all services and for every service it search for all connecting vertices (products) and checks if all of them contains in prods.
LET prods = (
FOR product IN products
FILTER product.ProductType == 'Product'
RETURN product
)
FOR service IN products
FILTER service.ProductType == 'Service'
FILTER prods ALL IN (
FOR v, e, p IN 1 OUTBOUND service edgeCol
RETURN v
)
RETURN service
With only 10 products and over 60,000 services you should think about splitting your collection products into two collection products and services where products only contains your 10 products. You can then drop the field ProductType. This allow more performant querying.
LET prods = (
FOR product IN products
RETURN product
)
FOR service IN services
FILTER prods ALL IN (
FOR v, e, p IN 1 OUTBOUND service edgeCol
RETURN v
)
RETURN service
