To lazily create the nodes use MERGE in the first line instead of MATCH
Am 21.02.2014 um 11:30 schrieb Ben Titmarsh <[email protected]>:
> Hi Michael,
>
> Based on your advice I changed my query to:
>
> MATCH (fromCard:Card { name: 'Hearts'}),(toCard:Card{ name: 'Spades'})
> CREATE UNIQUE (fromCard)-[r:DECK_INCIDENCE]-(toCard)
> SET r.passCount = coalesce(r.passCount, 0) +1
>
> I deleted one of the relationships and surely enough which ever order the
> cards are specified in it updates the relationship.
>
> However, it doesn't create the nodes 'Hearts' and 'Spades' if they don't yet
> exist. I realise that I didn't specify this in my previous post but I need
> the query to lazily create the nodes too, so in summary I need it to:
>
> 1) Lazily Create Nodes
> 2) Lazily Create Relationships
> 3) Increment the passCount property on the relationship whether it was just
> created or preexisted regardless of the direction of the relationship.
>
> I'll certainly leverage MERGE without arrows when it's released, but in the
> mean time I'd really like to get something live and working. Any guidance
> would be much appreciated!
>
> Ben.
>
> On Thursday, February 20, 2014 10:51:16 PM UTC, Michael Hunger wrote:
> MERGE without arrows were added and will be available in the next release.
>
> Until then try to replace that last merge with CREATE UNIQUE
>
> Michael
>
> Am 20.02.2014 um 23:30 schrieb Ben Titmarsh <[email protected]>:
>
>> I'd bluffed myself into thinking that I'd solved this but actually I hadn't.
>> This is my cypher:
>>
>> MATCH (fromCard:Card { name: {fromCardName}}),(toCard:Card{ name:
>> {toCardName}})
>> MERGE (fromCard:Card)-[r:DECK_INCIDENCE]->(toCard:Card)
>> SET r.passCount = coalesce(r.passCount, 0) +1
>>
>> For each pair of cards I end up creating a relationship in both directions,
>> for example the first time fromCard='HEARTS' and toCard='SPADES',
>> HEARTS->SPADES is created with passCount 1, then when fromCard='SPADES' and
>> toCard='HEARTS' SPADES->HEARTS is created with passCount 1. The desired
>> behaviour is that there is only one "bidirectional" relationship. The
>> direction is really not important, all I'm modelling is the strength of the
>> relationship i.e. the number of times that this combination of cards is
>> "passed" (note that I've used Hearts and Spades here purely for explanation,
>> I in fact have ~20k domain-specific cards)
>>
>> I tried without the arrow
>>
>> MATCH (fromCard:Card { name: {fromCardName}}),(toCard:Card{ name:
>> {toCardName}})
>> MERGE (fromCard:Card)-[r:DECK_INCIDENCE]-(toCard:Card)
>> SET r.passCount = coalesce(r.passCount, 0) +1
>>
>> but that gives me: Only directed relationships are supported in CREATE,
>> while MATCH allows to ignore direction.
>>
>> So I'm stuck here. It's important that relationships are created lazily,
>> i.e. the first time a 'pass' is registered, as for may pairs of cards they
>> will never be related at all. So how can I use the semantics of a merge to
>> create a relationship where the property passCount can be incremented (or
>> coalesced if null) regardless of the direction?
>>
>>
>>
>> On Wednesday, February 12, 2014 12:55:19 PM UTC, Peter Neubauer wrote:
>> Cool Ben.
>>
>> G: neubauer.peter
>> S: peter.neubauer
>> P: +46 704 106975
>> L: http://www.linkedin.com/in/neubauer
>> T: @peterneubauer
>>
>> Neo4j 2.0.0 - (graphs)-[:FOR]->(everyone)
>> Kids LAN creative party in Malmö - Kidscraft ICE
>>
>>
>> On Wed, Feb 12, 2014 at 1:45 PM, Ben Titmarsh <[email protected]> wrote:
>> > Ah my mistake! Didn't qualify my match with the property that I was
>> > matching on! This works:
>> >
>> > MATCH (fromCard:Card { name: {fromCardName}}),(toCard:Card{ name:
>> > {toCardName}})
>> >
>> > MERGE (fromCard:Card)-[r:DECK_INCIDENCE]->(toCard:Card)
>> > SET r.deckCount = coalesce(r.deckCount, 0) +1
>> >
>> > On Wednesday, February 12, 2014 11:56:32 AM UTC, Ben Titmarsh wrote:
>> >>
>> >> Hi Guys,
>> >>
>> >> I am aware that Neo4J doesn't have a concept of a bidirectional
>> >> relationship and I believe that's causing me a problem when using MERGE.
>> >> Here is my Cypher:
>> >>
>> >> MATCH (fromCard:Card),(toCard:Card)
>> >> MERGE (fromCard:Card)-[r:DECK_INCIDENCE]->(toCard:Card)
>> >> SET r.deckCount = coalesce(r.deckCount, 0) +1
>> >>
>> >> So say I have a collection of 4 cards:
>> >> "Hearts","Diamonds","Clubs","Spades"
>> >>
>> >> The first time I create the relationship between Hearts and Diamonds,
>> >> i.e.
>> >> Hearts -> Diamonds. However, the next time the cards are in a different
>> >> order and when I attempt to MERGE Diamonds -> Hearts a new relationship
>> >> is
>> >> created because no relationship exists in that direction. This is not
>> >> what
>> >> I want, I need to increment the deckCount property on the relationship
>> >> regardless of the direction and don't want to have to manage two
>> >> relationships per card.
>> >>
>> >> I cannot guarantee the order of cards and in essence I need to model a
>> >> Bidirectional Relationship. How can I achieve what I need using MERGE?
>> >>
>> >> Cheers,
>> >> Ben.
>> >
>> > --
>> > You received this message because you are subscribed to the Google Groups
>> > "Neo4j" group.
>> > To unsubscribe from this group and stop receiving emails from it, send an
>> > email to [email protected].
>> > For more options, visit https://groups.google.com/groups/opt_out.
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "Neo4j" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to [email protected].
>> For more options, visit https://groups.google.com/groups/opt_out.
>
>
> --
> You received this message because you are subscribed to the Google Groups
> "Neo4j" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> For more options, visit https://groups.google.com/groups/opt_out.
--
You received this message because you are subscribed to the Google Groups
"Neo4j" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.
