[tinkerpop 3.0.1incubating] simple mathematical operations ( + - / * ) after a .select('valA', 'valB',...)

[Florent B.]

Hi everyone,

I wondered how could I perform simple operations using gremlin groovy on selected elements.

Query example :

 

v1Id = 123456;
v2Id = 654321;

g.V(v1Id).as('v1').match(
  __.as('v1').both('someEdgeLabel').count().as('v1LinkedVertices').filter(
    both('someEdgeLabel').id().is(eq(v2Id))
  ).count().as('v1DirectCommonVerticesWithV2'),
).select('v1LinkedVertices', 'v1DirectCommonVerticesWithV2')

Then, I get this kind of result : 

[
  {
    v1LinkedVertices: x, // x = number
    v1DirectCommonVerticesWithV2: y // y = number
  }
]

How can I perform some basic mathematical operations so I can get this kind of result

[
  {
    v1LinkedVertices_MinusOne: x-1, // x = number
    v1DirectCommonVerticesWithV2_MultipliedByTwo: 2*y, // x = number
    stuffDoneHere: z // z = x/y

  }
]

 

Regards,

Flo.

[Daniel Kuppitz]

Your query doesn't look valid, hence let's play with ages in the modern graph.

gremlin> g.V().values("age")

==>29

==>27

==>32

==>35

Sum all ages:

gremlin> g.V().values("age").sum()

==>123

Multiply all ages:

gremlin> g.V().values("age").fold(1, mult)

==>876960

Subtract 1:

gremlin> g.V().values("age").map(union(identity(), constant(-1)).sum())

==>28

==>26

==>31

==>34

Multiply by 2 (simple):

gremlin> g.V().values("age").map(union(identity(), identity()).sum())

==>58

==>54

==>64

==>70

Multiply by 2 (generally useful for multiplications by n):

gremlin> g.V().values("age").map(union(identity(), constant(2)).fold(1, mult))

==>58

==>54

==>64

==>70

I don't have a solution for divisions that doesn't use lambdas.

Cheers,

Daniel

--
You received this message because you are subscribed to the Google Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/855b0fca-dae7-4a17-a6cf-634a078d54cb%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Florent B.]

Thanks (a lot) for your answer. I'll try to do some stuff with this.

I just hope it'll work with the select step too. :)

FYI, The query I presented was this query 'simplified' :

userId = 40968224
jobId = 12320
g.V().hasLabel('joboffer').as('job').match(
  __.as('job').out('tagged').count().as('jobTags'),
  __.as('job').both('tagged').hasLabel('tag').filter(
    both('knows').hasLabel('user').id().is(eq(userId))
  ).count().as('commonTags'),
).select('jobTags', 'commonTags')

I can assure you that this query is working, here is the result :

[ { jobTags: 5, commonTags: 3 } ]

:)

[Daniel Kuppitz]

userId = 40968224
jobId = 12320
g.V().hasLabel('joboffer').as('job').match(
  __.as('job').out('tagged').count().as('jobTags'),
  __.as('job').both('tagged').hasLabel('tag').filter(
    both('knows').hasLabel('user').id().is(eq(userId))
  ).count().as('commonTags'),
).select('jobTags', 'commonTags') 

I can assure you that this query is working

I have no complaints when I look at this query, but the first one was not simplified, it was just busted and invalid.

Cheers,

Daniel

To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/ce35a666-c788-47ee-a39e-00f5a2ccd79f%40googlegroups.com.

[Florent B.]

My bad ! :)

[Florent B.]

I don't have a solution for divisions that doesn't use lambdas.

 .fold(1){a,b -> b/a}

;)

Thanks a lot !

[Daniel Kuppitz]

FYI: That's a lambda. You can solve any use-case using lambdas, but lambda-less traversals are usually preferred as they perform a lot better.

Cheers,

Daniel

--

You received this message because you are subscribed to the Google Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-user...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/a2bacdca-f495-42da-97f8-df7df9e557fb%40googlegroups.com.

[Florent B.]

Good to know :)

Thanks again a lot for your help ! :)

[Manjunath Sindagi]

What is identity here? Instead of identity or constant values, is there a way to pass another query and run the mathematical operations?

[Daniel Kuppitz]

identity() is the current element / value, in this case the age.

Cheers,

Daniel

To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-users+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/183231ab-e69f-42e6-830a-4bc2e069d023%40googlegroups.com.

[Manjunath Sindagi]

Thanks Daniel.

Is there a way to fire two gremlin queries in one and do mathematical operation. ?

[Daniel Kuppitz]

Not sure what you're after. Can you provide a tiny sample graph and briefly describe what you want to do?

Cheers,

Daniel

To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-users+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/f7293ffd-d9fa-46bd-aef0-6e151e7350ff%40googlegroups.com.

[Manjunath Sindagi]

Daniel,

It's little difficult to explain over mail, i shall try though

Attaching the graph,

1. An input is given to like Say p1, we need to find the total count of has_dependencies. (here total Cs)

2. after findling total Cs, find the total has_dependencies Cs has.

The ratio of 1/2 needs to be found.

Independently 2 queries are fired. However, now figuring out if this whole 1 & 2 step can be merged in to 1.

Thanks,
Manjunath

[Daniel Kuppitz]

Hi Manunath,

I still don't get it. This is your sample graph:

g = TinkerGraph.open().traversal()
g.addV().property(id, 'A').as('a').
  addV().property(id, 'B1').as('b1').
  addV().property(id, 'B2').as('b2').
  addV().property(id, 'C').as('c').
  addV().property(id, 'D').as('d').
  addV().property(id, 'P1').as('p1').
  addV().property(id, 'P2').as('p2').
  addV().property(id, 'P3').as('p3').
  addE('has_info').from('a').to('b1').
  addE('has_info').from('a').to('b2').
  addE('has_dependency').from('c').to('b1').
  addE('has_dependency').from('d').to('b2').
  addE('xxx').from('b1').to('p1').
  addE('xxx').from('b1').to('p2').
  addE('xxx').from('b1').to('p3').iterate()

It's pretty clear how to get from P1 to C, but where do you get the 1:2 ratio from?

Cheers,

Daniel

To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-users+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/dfcfe21f-ec3d-4d51-9ae6-478b5587715b%40googlegroups.com.