Find siblings including self

[Benja1972]

I would like to create a simple query to the graph with hierarchy. Starting from a child I would like to find a parent and all children of this parent including starting node. Here the example of the code,

from grandcypher import GrandCypher
import networkx as nx


host = nx.DiGraph()
host.add_node("p",  foo="Parent")
host.add_node("c1", foo="Child01")
host.add_node("c2", foo="Child02")
host.add_node("c3", foo="Child03")
host.add_node("c4", foo="Child04")
host.add_node("c5", foo="Child05")


host.add_edge("c1", "p")
host.add_edge("c2", "p")
host.add_edge("c3", "p")
host.add_edge("c4", "p")
host.add_edge("c5", "p")


qry = """
MATCH (n {foo: "Child01"})-->()<--(s)
RETURN s.foo
"""
print(GrandCypher(host).run(qry))

In current implementation the code output exclude initial node Child01

{'s.foo': ['Child02', 'Child03', 'Child04', 'Child05']}

Is it implemented on purpose?
How I can alter this behavior in single query?

[j6k4m8]

This is interesting — I think by default we assume that entities in the result should be distinct, but I'm actually not sure how the Cypher reference implementation treats this (in Neo4j).

Is there a particular reason you're using Cypher to do this? This search query might be better suited to something like a nx.descendants or similar call!

[Benja1972]

Thank you for reply. I could manage to do this in networkx. The idea is to define a set of Cypher queries agnostic to the graph storage. The project, I am involved, could go from POC to prod and networkx could not fit as backend storage. The persistent graph database would be used. Moreover in my real example, from initial node I need to go up one level to the parent and retrieve branch to some level of depth:

qry = """
MATCH (n {foo: "Child01"})-->()<-[*0..3]-(s)
RETURN s.foo
"""