Moses
Moses

Reputation: 113

Get all edges between multiple vertices - JanusGraph

I am trying to write a gremlin query to get all the edges between a list of vertices.

Data redacted to protect privacy. User A(42651832) -reports_to-> User B(28729440) -reports_to-> User C(19546208)

ids = [19546208, 28729440, 42651832]

I need to find all the edges between an arbitrary list of vertices

Seems simple, but I am unable to write a query that gives results I need.

gremlin> g.V(42651832).outE('reports_to').otherV().id()
==>28729440
gremlin> g.V(28729440).outE('reports_to').otherV().id()
==>19546208
gremlin> ids = [19546208, 28729440, 42651832]
==>19546208
==>28729440
==>42651832
gremlin> g.V(ids)
==>v[19546208]
==>v[28729440]
==>v[42651832]
gremlin> g.V(ids).bothE().where(otherV().hasId(ids))
gremlin> g.V(ids).bothE().where(otherV().hasId(within(ids)))
gremlin> g.V(ids).bothE().where(otherV().hasId(within(19546208, 28729440, 42651832)))

Apparently, I think there is a type conversion issue between gremlin console and JanusGraph.

quoting as string or converting to Long seems to work.

gremlin> g.V(ids).bothE().where(otherV().hasId(within("19546208", "28729440", "42651832")))
==>e[10r7d8-h3rs0-i6t-bmxy8][28729440-reports_to->19546208]
==>e[10r7d8-h3rs0-i6t-bmxy8][28729440-reports_to->19546208]
==>e[128qvr-pe6d4-i6t-h3rs0][42651832-reports_to->28729440]
==>e[128qvr-pe6d4-i6t-h3rs0][42651832-reports_to->28729440]
gremlin> g.V(ids).bothE().where(otherV().hasId(within(19546208L, 28729440L, 42651832L)))
==>e[10r7d8-h3rs0-i6t-bmxy8][28729440-reports_to->19546208]
==>e[10r7d8-h3rs0-i6t-bmxy8][28729440-reports_to->19546208]
==>e[128qvr-pe6d4-i6t-h3rs0][42651832-reports_to->28729440]
==>e[128qvr-pe6d4-i6t-h3rs0][42651832-reports_to->28729440]
gremlin>  

gremlin> g.V(ids).bothE().where(otherV().hasId(within(19546208L, 28729440L, 42651832L))).dedup()
==>e[10r7d8-h3rs0-i6t-bmxy8][28729440-reports_to->19546208]
==>e[128qvr-pe6d4-i6t-h3rs0][42651832-reports_to->28729440]

Any other suggestions. Not sure why JanusGraph works this way.

Upvotes: 0

Views: 636

Answers (1)

Chris Hupman
Chris Hupman

Reputation: 408

This isn't quite a full answer, but hopefully it gets you close enough. I used GraphOfTheGods to test it out.

This will get all the paths from the ids list to the quoted ids in hasID() and then output a list of all the edges traversed in each path. I added a limit for readability. You could easily add all the values to a set to get a deduped answer.

# Save all the graph of the gods vertex ids to a variable
ids = [4112,4128,4136,8232,12328,16424,20520,4296,4328,4344,8440,12536]
paths = g.V(ids).until(hasId("8440","12536")).repeat(bothE().aggregate("e").otherV().simplePath()).limit(3).select('e')
==>[e[74v-6ig-9hx-368][8440-battled->4112]]
==>[e[74v-6ig-9hx-368][8440-battled->4112],e[7xb-6ig-9hx-36o][8440-battled->4128],e[1l0-36o-b2t-9o8][4128-lives->12536],e[9vp-co8-bv9-36o][16424-pet->4128]]
==>[e[74v-6ig-9hx-368][8440-battled->4112],e[7xb-6ig-9hx-36o][8440-battled->4128],e[1l0-36o-b2t-9o8][4128-lives->12536],e[9vp-co8-bv9-36o][16424-pet->4128]]

What I was able to originally get was a full path with connecting vertexes which I am including in case it could potentially be helpful.

paths = g.V(ids).until(hasId("8440","12536")).repeat(bothE().otherV().simplePath()).path().limit(5)
==>[v[4112],e[74v-6ig-9hx-368][8440-battled->4112],v[8440]]
==>[v[4128],e[7xb-6ig-9hx-36o][8440-battled->4128],v[8440]]
==>[v[4128],e[1l0-36o-b2t-9o8][4128-lives->12536],v[12536]]
==>[v[4128],e[9vp-co8-bv9-36o][16424-pet->4128],v[16424],e[9hh-co8-b2t-9o8][16424-lives->12536],v[12536]]
==>[v[4128],e[9vp-co8-bv9-36o][16424-pet->4128],v[16424],e[8p1-co8-cnp-3co][16424-brother->4344],v[4344],e[6cf-6ig-7x1-3co][8440-father->4344],v[8440]]

Separately I did some checking with the GraphOfTheGods and the explain() step and it definitely seems like a bug. If I set a list to variable it is performing an equal step instead of a within step.

paths = g.V(ids).until(hasId(ids)).repeat(out().simplePath()).limit(10).path().explain()
...RepeatStep(until([HasStep([~id.eq([4112, 4128, ...])])]),

where as listing in quotes it will properly do a within check.

paths = g.V(ids).until(hasId("8440","12536")).repeat(outE().simplePath()).limit(10).path().explain()
...RepeatStep(until([HasStep([~id.within([8440, 12536])])])

Upvotes: 1

Related Questions