Apache TinkerPop•2y ago

Query if else in gremlin

I have a query and that return a list vertex. I want to do a query from those vertexes like this (if - else): - vertex doesn't has out edge -> return itself - vertex has out edge -> then keep querying until there are no edges out ( repeat(out()).until(outE().count().is(0)) ) I'm trying to limit the use of loops here to improve performance. Who can help me. Thank you very much

7 Replies

spmallette•2y ago

The functionality you are describing is basically what repeat() does by default, but you have to tell it to actually "return itself" with emit(). Here's an example:

gremlin> g = TinkerFactory.createModern().traversal()
==>graphtraversalsource[tinkergraph[vertices:6 edges:6], standard]
gremlin> g.V(1).repeat(out())
gremlin>

gremlin> g = TinkerFactory.createModern().traversal()
==>graphtraversalsource[tinkergraph[vertices:6 edges:6], standard]
gremlin> g.V(1).repeat(out())
gremlin>

The above traverses out() from v[1] until it runs out of edges to follow. If you add emit() (which takes a predicate and defaults to true when none is supplied) you tell repeat() under what circumstance that you want to actually put those vertices into the stream, so:

gremlin> g.V(1).repeat(out()).emit()
==>v[3]
==>v[2]
==>v[4]
==>v[5]
==>v[3]

gremlin> g.V(1).repeat(out()).emit()
==>v[3]
==>v[2]
==>v[4]
==>v[5]
==>v[3]

In your case, you really don't want all of those vertices because v[4] doesn't match your criteria as it has outgoing edges. You can prevent it from being emitted like:

gremlin> g.V(1).repeat(out()).emit(__.not(outE()))
==>v[3]
==>v[2]
==>v[5]
==>v[3]

gremlin> g.V(1).repeat(out()).emit(__.not(outE()))
==>v[3]
==>v[2]
==>v[5]
==>v[3]

There are some duplicates in the case of the "modern" graph, you could add dedup() to the end to polish that up:

gremlin> g.V(1).repeat(out()).emit(__.not(outE())).dedup()
==>v[3]
==>v[2]
==>v[5]

gremlin> g.V(1).repeat(out()).emit(__.not(outE())).dedup()
==>v[3]
==>v[2]
==>v[5]

Dinh PhuOP•2y ago

Can i ask another question? When i run this below query: "g.V().has("guid", "6203620951906330066").limit(10).out("matching").profile().toSet()" And I see step "JanusGraphVertexStep(OUT,[matching],vertex)" take a lot of time. Is there any way to solve this?

query.txt

spmallette•2y ago

i think you should ask a separate question for that. that way janusgraph folks might have a better chance of spotting it. i don't know the answer offhand.

Dinh PhuOP•2y ago

ok, thank you

Dinh PhuOP•2y ago

In your query

g.V(1).repeat(out()).emit(__.not(outE())).dedup()

g.V(1).repeat(out()).emit(__.not(outE())).dedup()

. If V(1) doesn't have out edge, it will not return V(1). While i want return itself if it doesn't have out edge

spmallette•2y ago

you could use optional() in this case:

g.V(1).optional(repeat(out()).emit(__.not(outE())).dedup())

g.V(1).optional(repeat(out()).emit(__.not(outE())).dedup())

Dinh PhuOP•2y ago

oh, thank you

Gaming

Programming

Query if else in gremlin