Just like you can dynamically WHERE on the type of a relationship, I
thought there was a way to dynamically WHERE on the direction as well -- by
comparing the start and end nodes of a relationship.
But neither START(rel) / END(rel) nor rel.start / rel.end work for me, and
neither on 1.6 nor on 1.8/9 (whatever's on console.neo4j.org). And looking
through the docs, I can't find any mention of these functions.
Is this not implemented yet and my memory was only of a discussion /
feature request? Or is it some other name? Thanks!
> Just like you can dynamically WHERE on the type of a relationship, I
> thought there was a way to dynamically WHERE on the direction as well -- by
> comparing the start and end nodes of a relationship.
> But neither START(rel) / END(rel) nor rel.start / rel.end work for me, and
> neither on 1.6 nor on 1.8/9 (whatever's on console.neo4j.org). And
> looking through the docs, I can't find any mention of these functions.
> Is this not implemented yet and my memory was only of a discussion /
> feature request? Or is it some other name? Thanks!
> Just like you can dynamically WHERE on the type of a relationship, I thought there was a way to dynamically WHERE on the direction as well -- by comparing the start and end nodes of a relationship.
> But neither START(rel) / END(rel) nor rel.start / rel.end work for me, and neither on 1.6 nor on 1.8/9 (whatever's on console.neo4j.org). And looking through the docs, I can't find any mention of these functions.
> Is this not implemented yet and my memory was only of a discussion / feature request? Or is it some other name? Thanks!
> Also treating relationships as a short path would be a nice thing.
> so one can do nodes(r). or head(nodes(r)) tail(nodes(r)).
> Michael
> Am 13.10.2012 um 16:43 schrieb Aseem Kishore:
> Just like you can dynamically WHERE on the type of a relationship, I
> thought there was a way to dynamically WHERE on the direction as well -- by
> comparing the start and end nodes of a relationship.
> But neither START(rel) / END(rel) nor rel.start / rel.end work for me, and
> neither on 1.6 nor on 1.8/9 (whatever's on console.neo4j.org). And
> looking through the docs, I can't find any mention of these functions.
> Is this not implemented yet and my memory was only of a discussion /
> feature request? Or is it some other name? Thanks!
> Also treating relationships as a short path would be a nice thing.
> so one can do nodes(r). or head(nodes(r)) tail(nodes(r)).
> Michael
> Am 13.10.2012 um 16:43 schrieb Aseem Kishore:
>> Just like you can dynamically WHERE on the type of a relationship, I thought there was a way to dynamically WHERE on the direction as well -- by comparing the start and end nodes of a relationship.
>> But neither START(rel) / END(rel) nor rel.start / rel.end work for me, and neither on 1.6 nor on 1.8/9 (whatever's on console.neo4j.org). And looking through the docs, I can't find any mention of these functions.
>> Is this not implemented yet and my memory was only of a discussion / feature request? Or is it some other name? Thanks!
Ah, in my case at least, I'm interested in both directions -- hence not
specifying a direction in MATCH. I'm having trouble imagining a case where
you *wouldn't* be interested in both directions if you wanted to WHERE on
the direction.
(In my case, my WHERE looks something like (direction in AND foo) OR
(direction out AND bar).)
Aseem
On Sat, Oct 13, 2012 at 12:13 PM, Michael Hunger <
> Ah, in my case at least, I'm interested in both directions -- hence not
> specifying a direction in MATCH. I'm having trouble imagining a case where
> you *wouldn't* be interested in both directions if you wanted to WHERE on
> the direction.
> (In my case, my WHERE looks something like (direction in AND foo) OR
> (direction out AND bar).)
> Aseem
> On Sat, Oct 13, 2012 at 12:13 PM, Michael Hunger <
> michael.hun...@neotechnology.com> wrote:
>> Sure I know :)
>> Was just demonstrating a "true" case
>> Am 13.10.2012 um 17:48 schrieb Andres Taylor:
>> On Sat, Oct 13, 2012 at 5:43 PM, Michael Hunger <
>> michael.hun...@neotechnology.com> wrote:
>>> actually you can even use
>>> MATCH p = (a) -[r:FOLLOWS]- (b)
>>> WHERE a = head(nodes(p))
>> Sorry amigo, but that doesn't work. The path always follows the pattern
>> you have described - in this pattern, a will _always_ be a the head of p.
> On Sat, Oct 13, 2012 at 8:35 PM, Aseem Kishore <aseem.kish...@gmail.com>
wrote:
>> Ah, in my case at least, I'm interested in both directions -- hence not
specifying a direction in MATCH. I'm having trouble imagining a case where
you *wouldn't* be interested in both directions if you wanted to WHERE on
the direction.
>> (In my case, my WHERE looks something like (direction in AND foo) OR