Hi Craig, thanks, yes i know i could have combined my query into one but this was just to split things out. Its only purpose was to create the initial subset of the bigger graph that would be input for the actual query.. I tried your query but it does not give any results back but with some help from stackoverflow i learned that the ALL clause did the trick..
A working version of the cypher query ( i.e. a simplified version for only looking at the attribute relationship and not yet at the policyclass relation) is : MATCH (u:User)-[:ASSIGNED_TO]->(ua:Attribute)-[:ASSOCIATED_TO]->(oa:Attribute) WHERE u.name = "u1" WITH u, COLLECT(oa) AS Attributes MATCH (o:Object) WHERE ALL( x IN Attributes WHERE (o)-[*]->(x) ) RETURN u, Collect(o) AS Objects Thanks for you reply ! regards koen 2018-01-04 14:23 GMT+01:00 Craig Taverner <cr...@amanzi.com>: > Hi Koen, > > As I see it one problem with your query is that it is performing multiple > matches that will build on each other, and that is not what you are asking > for. The first match will return two rows, one for each of the two paths > that match that pattern. Then for each of these rows Cypher will run the > second match, leading to four rows (with duplications), and so on. There is > nothing in the query to say that you want a correlation between the first > match and any other match. > > One simple change that might be closer to what you want would be to remove > all the intermediate WITH *, and MATCH statements, and make a single, > comma-separated MATCH, like this: > > MATCH p = (u:User)-[:ASSIGNED_TO*]->(pc:PolicyClass), > (u)-[:ASSIGNED_TO*]->(:Attribute)-[:ASSOCIATED_TO]->(a1:Attribute), > (o:Object)-[:ASSIGNED_TO*]->(pc), > (o)-[:ASSIGNED_TO*]->(a1), > RETURN p > > This single MATCH clause now describes a single complete pattern, which > should be matched twice only. Try this and perhaps simplified versions of > it, and see if that works. > > Regards, Craig > > On Wednesday, January 3, 2018 at 11:08:24 AM UTC+1, koen wrote: >> >> Hi, a cypher question. How can you check if 2 (or more) different nodes >> in a graph have relationships to multiple other (shared) nodes in the >> graph. >> >> Attached is a graph (the result of below cypher query) that i created >> based on the sample provided in the NIST / INCITS NGAC 499 / 525 (Next >> Generation Access Control) specification (graph.png). I also included the >> diagram of the complete sample graph. >> >> BLUE --- : u1 is a :User node >> GREEN --- a11, a21, l11, l12 are :Object nodes >> YELLOW --- branch-constrants, position-constraints are :PolicyClass nodes >> GREY --- all other nodes are attribute nodes >> >> In the sample provided "u1" has a relationship with the 2 policy classes >> (branch and position constraints) and it has an ASSOCIATED_TO relation ship >> to both "products1" and "accounts" Attribute Nodes >> Looking from the :Object side only "a11" has a relationship to both the >> 2 PolicyClasses and to the "products1" and "accounts" attribute nodes -- >> simular to what u1 has "u1" >> >> All other Objects (a21, l12 and l11) do not have these 4 relation ships >> but only some of these (meaning u1 has no acces to these objects) >> >> How can i further filter this out in cypher ? So i would like to further >> refine the result so that only "u1" and "a11" (meaning user u1 has access >> to object a11) are part of the end result set. >> >> it seems straightforward but it looks like i am running on some >> constraints where some nodes are visited only once and no results are >> returned (for example going from "a11" you have to visit "accounts1" node >> multiple times) >> >> This is the cypher code used to create the graph as speficied in attached >> diagram. >> >> MATCH p1 = (u:User)-[:ASSIGNED_TO*]->(pc:PolicyClass) >> WHERE u.name = "u1" >> WITH * >> MATCH p2 = (u)-[:ASSIGNED_TO*]->(a:Attribute)-[:ASSOCIATED_TO]->(a1: >> Attribute) >> WITH * >> MATCH p3 = (o:Object)-[:ASSIGNED_TO*]->(pc:PolicyClass) >> WITH * >> MATCH p4 = (o)-[:ASSIGNED_TO*]->(a1:Attribute) >> WITH * >> RETURN p1,p2,p3,p4 >> >> Any help is welcome ! >> >> Thanks regards Koen >> > -- > You received this message because you are subscribed to a topic in the > Google Groups "Neo4j" group. > To unsubscribe from this topic, visit https://groups.google.com/d/ > topic/neo4j/94ULtuII4g4/unsubscribe. > To unsubscribe from this group and all its topics, send an email to > neo4j+unsubscr...@googlegroups.com. > For more options, visit https://groups.google.com/d/optout. > -- Groeten / Best Regards, Koen -- 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 neo4j+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
