Reputation: 13
The neo4j Graph Data Science (GDS) "Triangle Count" algorithm only runs on an undirected graph. I can convert my existing directed graph to undirected for the calculation using:
CALL gds.graph.create('myGraph3', '*', {BELONGS_TO: {orientation: 'UNDIRECTED'}, FLOWS_TO: {orientation: 'UNDIRECTED'}})
But I have to list every relationship, which is brittle and lengthy for the actual list of relationships. Is there a way to make the '*' work with this?
Something like
CALL gds.graph.create('myGraph3', '', {'': {orientation: 'UNDIRECTED'}})
which does not work, but I can't figure the correct pattern.
Upvotes: 0
Views: 887
Reputation: 1
You can create a ephemeral graph from the projection. This ephemeral graph is an in-memory version that can be queried with Cypher so that you can see if it is correct or not.
CALL gds.ephemeral.database.create( 'ephemeralDatabaseName', 'existingProjectedGraph' ) YIELD dbName, graphName, createMillis
Upvotes: 0
Reputation: 6534
Yes, that is possible, there is a type parameter where you can input the wildcard option:
CALL gds.graph.create('myGraph3',
'*',
{ALL: {orientation: 'UNDIRECTED', type:'*'}})
Upvotes: 0