How to map the start vertex to all the vertices it is reachable to? - gremlin

Question

I am trying to come up with a gremlin query that maps the start node to all the nodes it is reachable to.
So for example, if the graph is something like this:
node1--link-->node2--link-->node3--link-->node4
the response I am expecting from query is:
[start=node1, relates=[node2, node3, node4]]
So here, if node3 is reachable to lets say node5 and node6, they also should be included in the result.
I have written the query: https://gremlify.com/hatp1roeii
But the response is not what I expected.

Answer 1

Using the test data (reformatted for readability) from the Gremlify example:
g.addV('node').as('1').property(single, 'recordId', 'node9').property(single, 'data', 22).
addV('node').as('2').property(single, 'recordId', 'node10').property(single, 'data', 16).
addV('root').as('3').property(single, 'recordId', 'node1').property(single, 'data', 9).
addV('node').as('4').property(single, 'recordId', 'node2').property(single, 'data', 5).
addV('node').as('5').property(single, 'recordId', 'node4').property(single, 'data', 2).
addV('node').as('6').property(single, 'recordId', 'node3').property(single, 'data', 11).
addV('node').as('7').property(single, 'recordId', 'node7').property(single, 'data', 15).
addV('node').as('8').property(single, 'recordId', 'node6').property(single, 'data', 10).
addV('node').as('9').property(single, 'recordId', 'node8').property(single, 'data', 1).
addV('node').as('10').property(single, 'recordId', 'node5').property(single, 'data', 8).
addE('left').from('1').to('2').
addE('left').from('3').to('4').
addE('right').from('3').to('6').
addE('left').from('4').to('5').
addE('right').from('4').to('10').
addE('left').from('5').to('9').
addE('left').from('6').to('8').
addE('right').from('6').to('7').
addE('right').from('7').to('1')
All of the paths from 'node1' can be discovered using
gremlin> g.V().has('recordId','node1').
......1> repeat(out()).
......2> until(__.not(out())).
......3> path().by('recordId')
==>[node1,node2,node5]
==>[node1,node3,node6]
==>[node1,node2,node4,node8]
==>[node1,node3,node7,node9,node10]
We can dedup that result to get the unique list of nodes visited
gremlin> g.V().has('recordId','node1').
......1> repeat(out()).
......2> until(__.not(out())).
......3> path().by('recordId').
......4> unfold().
......5> dedup().
......6> fold()
==>[node1,node2,node5,node3,node6,node4,node8,node7,node9,node10]
While not the nicely formatted answer yet, this is essentially one way of solving the problem. We can modify the query above to get the formatted result described in the question:
gremlin> g.V().has('recordId','node1').
......1> repeat(out()).
......2> until(__.not(out())).
......3> path().by('recordId').
......4> unfold().
......5> dedup().
......6> fold().
......7> project('start','relates_to').
......8> by(limit(local,1)).
......9> by(skip(local,1))
==>[start:node1,relates_to:[node2,node5,node3,node6,node4,node8,node7,node9,node10]]
Here is the same query, but starting at 'node7'
gremlin> g.V().has('recordId','node7').
......1> repeat(out()).
......2> until(__.not(out())).
......3> path().by('recordId').
......4> unfold().
......5> dedup().
......6> fold().
......7> project('start','relates_to').
......8> by(limit(local,1)).
......9> by(skip(local,1))
==>[start:node7,relates_to:[node9,node10]]
Lastly, here's a version that does not use the path step, but instead uses store to aggregate the nodes visited.
gremlin> g.V().has('recordId','node1').store('start').
......1> repeat(out().store('seen')).
......2> until(__.not(out())).
......3> cap('seen').dedup(local).unfold().values('recordId').fold().
......4> project('start','relates_to').
......5> by(cap('start').unfold().values('recordId')).
......6> by()
==>[start:node1,relates_to:[node2,node3,node4,node5,node6,node7,node8,node9,node10]]
and from 'node7'
gremlin> g.V().has('recordId','node7').store('start').
......1> repeat(out().store('seen')).
......2> until(__.not(out())).
......3> cap('seen').dedup(local).unfold().values('recordId').fold().
......4> project('start','relates_to').
......5> by(cap('start').unfold().values('recordId')).
......6> by()
==>[start:node7,relates_to:[node9,node10]]