Hey everyone,

I have this really cool application for which I think Neo4J would be an 
excellent solution, but there are still some issues I'd need to figure out. 
I posted my question on stackoverflow but thought I might try my luck in 
the google group as well.

I'm trying to create a query that gives me the smallest set of article 
numbers that fully describes the configuration of a customer order and not 
more than that. 

In my graph I have two types of nodes: FeatureValues and Articles. One 
article can provide one or more FeatureValues. My input is a set of 
FeatureValues and I need to know which is the smallest set of article 
numbers that completely describes all my input values and doesn't add any 
additional FeatureValues. Best case scenario is that the requested 
configuration can be covered with one article number, but it's also 
possible that more than one article number is required:

Example dataset:

FV1:FeatureValue
FV2:FeatureValue
FV3:FeatureValue
FV4:FeatureValue

A1:Article-[:PROVIDES]->FV1
A2:Article-[:PROVIDES]->FV2
A3:Article-[:PROVIDES]->FV3
A4:Article-[:PROVIDES]->FV4

A5:Article-[:PROVIDES]->FV1 and FV2
A6:Article-[:PROVIDES]->FV1 and FV2 and FV3

Possible cases:

input = FV1, FV2
output: A5

input = FV1, FV2, FV3
output: A6

input = FV1, FV2, FV4
output: A4, A5
 

I found a somewhat related issue here: Neo4J - find nodes where the nodes 
related are a subset 
<https://stackoverflow.com/questions/27376831/neo4j-find-nodes-where-the-nodes-related-are-a-subset>.
 
The result of that query is a list of all item numbers that provide the 
requested FeatureValues and don't add anything else, but it's not the 
smallest subset as some FeatureValues will be described by more than one 
article number. Is what I want to do feasible?

Thanks!

-- 
You received this message because you are subscribed to the Google Groups 
"Neo4j" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to