CODEWITHSUNDEEP
gremlin
jiamo
jiamo

Reputation: 1456

Bool logic (or) in gremlin match?

In gremlin I have such query:

g.V().match(
  __.as('a').outE('x').as('a_x_b').inV().as('b'),
  __.as('a').outE('x1').as('a_x1_b').inV().as('b'),
  __.as('a').outE('x2').as('a_x2_b').inV().as('b'),
  __.as('b').outE('know').as('b_knows_c').inV().as('c'),
  __.as('b').outE('know1').as('b_knows1_c').inV().as('c')
).select('a','b','c', 'a_x_b', 'a_x1_b', 'b_knows1_c', 'a_x2_b')

But it is and logic. I want to using or in g such like

g.V().match(
  or(__.as('a').outE('x').as('a_x_b').inV().as('b'),
     __.as('a').outE('x1').as('a_x1_b').inV().as('b'),
     __.as('a').outE('x2').as('a_x2_b').inV().as('b')),
  or(__.as('b').outE('know').as('b_knows_c').inV().as('c'),
     __.as('b').outE('know1').as('b_knows1_c').inV().as('c'))
).select('a','b','c', 'a_x_b', 'a_x1_b', 'b_knows1_c', 'a_x2_b')

I think there is no such thing at gremlin. Why I need such thing. Because the one property of edge don't support multi value like the vertex. I want such multi value in one edge, So I must using multi label. Then I need to filter such label edge but using the or logic.

In another way: how to choose some edge label from multi label?

I figure out to such answer.

g.V().match(
  __.as('a').outE().as('a_x_b').filter(or(hasLabel('x'), hasLabel('x2'))).inV().as('b'),    
  __.as('b').outE().as('b_know_c').filter(or(hasLabel('know'), hasLabel('know1'))).inV().as('c')
).select('a','b','c', 'a_x_b', 'b_know_c')

still open the question for more methods.

turn label to property is ok too.

g.V().match(
  __.as('a').outE('x4').as('a_x_b').where(or(has('tag1', 1), has('tag2', 1))).inV().as('b'),
  __.as('b').outE().as('b_know_c').filter(or(hasLabel('know'), hasLabel('know1'))).inV().as('c')).select('a','b','c', 'a_x_b', 'b_know_c')

Upvotes: 0

Views: 128

Answers (1)

Kelvin Lawrence
Kelvin Lawrence

Reputation: 14371

A few observations:

  1. In general I try to avoid the match step. It often leads to inefficient query execution. Pretty much everything you can do with a match can also be done using where steps.
  2. You don't really need to use any or logic here. You can provide multiple labels to steps like outE and hasLabel. For example:
gremlin> g.V(44).inE().label()
==>contains
==>contains
==>route
==>route
==>route
==>route
==>extra

gremlin> g.V(44).inE('extra','contains')
==>e[61286][8-extra->44]
==>e[57826][3743-contains->44]
==>e[54323][3729-contains->44]

gremlin> g.V(44).inE().hasLabel('extra','contains')
==>e[61286][8-extra->44]
==>e[57826][3743-contains->44]
==>e[54323][3729-contains->44]

Upvotes: 1

Related Questions