Count incoming and outgoing edges from a given vertex to each neighbour in a multidigraph

Question

I am trying to find a query that would allow me to get a sort of group count of incoming and outgoing vertices of a vertex in a multidigraph.

Taking the graph above for V[0], we should obtain :

[V[0],V[1],incoming:2,outgoing:0]

[V[0],V[2],incoming:1,outgoing:0]

[V[0],V[3],incoming:0,outgoing:1]

Answer 1

When asking questions about Gremlin, a picture and graph description are nice, but a Gremlin script creating some sample data is even better:

g = TinkerGraph.open().traversal()
g.addV('node').property(T.id,0).as('0').
  addV('node').property(T.id,1).as('1').
  addV('node').property(T.id,2).as('2').
  addV('node').property(T.id,3).as('3').
  addE('link').from('1').to('0').
  addE('link').from('1').to('0').
  addE('link').from('0').to('3').
  addE('link').from('2').to('0').iterate()

Here's one way to do this:

gremlin> g.V(0).bothE().
......1>   group().
......2>     by(union(inV(),outV()).fold()).
......3>     by(fold().
......4>        project('incoming','outgoing').
......5>          by(unfold().inV().hasId(0).count()).
......6>          by(unfold().outV().hasId(0).count()))
==>[[v[0],v[1]]:[incoming:2,outgoing:0],[v[0],v[2]]:[incoming:1,outgoing:0],[v[3],v[0]]:[incoming:0,outgoing:1]]

Basically, we group() each edge by its associated in/out vertices (line 2 - i.e. a List of in/out formed by union().fold()) and then reduce the edges that are gathered per vertex pair (starting at line 3). The reduce operation simply creates a list with fold() then uses project() to transform that List to a Map with "incoming" and "outgoing" keys - the values for those keys are defined in the following respective by() modulators (i.e. unfold the list of edges, filter for vertex "0" appropriately and count()).