gremlin add vertex between two vertex - in a single query?

Question

Let's say I have a social graph and part of it represents people attending parties and an app for tracking invitations. I have a vertex for John, and a vertex for David, who both attended the same party.

In the app, John says he invited Sarah, and David says he was invited by Sarah, but I don't have a vertex for Sarah.

In a single query, can I create Sarah and connect John to Sarah and Sarah to David?

So far, the best I can do is determine that Sarah is missing from people with the following, using the coalesce pattern for get or create:

g.V().has('name', 'John').out().has('name', 'Sarah').as_('x').out().has('name', 'David').select('x').fold().coalesce(unfold(), constant('how to I put Sarah between John and David here?')).next()

Can it all be done in one query - replacing my constant with a multi step mutating traversal?

EDIT:

I got a bit further with it. I think the following works if I already have a vertex for Sarah: g.V().has('name', 'John').addE('invited').to(g.V().has('name', 'Sarah').as_('x').addE('invited').to(g.V().has('name', 'David')).select('x')).next()

Is this the recommended approach? Is there a way to create the Sarah vertex in the same query? Would I nest another .fold().coalesce(unfold(), ...) in there?

Answer 1

Checking for Sarah first and creating if not found may simplify your query. This is from the Gremlin Console.

gremlin> g.addV('person').property('name','John')
==>v[62869]
gremlin> g.addV('person').property('name','David')
==>v[62871]    

gremlin> g.V().has('name','Sarah').
......1>   fold().
......2>   coalesce(unfold(),
......3>             addV('person').property('name','Sarah')).as('s').
......4>   addE('invited').to(V().has('name','David')).
......5>   addE('invited').from(V().has('name','John')).to('s')    
==>e[62876][62869-invited->62873] 

gremlin> g.V().has('name','Sarah').outE().inV().path().by('name').by(label)
==>[Sarah,invited,David]
gremlin> g.V().has('name','Sarah').inE().outV().path().by('name').by(label)
==>[Sarah,invited,John]