The problem is just exploding cardinality. Each new mach multiplies the
cardinality of the (intermediate) results
Your first solution is a good one
Another one that you might want to do is:
MATCH (a:A{email:"[email protected]"})
OPTIONAL MATCH (a)-[r1:r1]-()
with a, collect(r1) as r1
OPTIONAL MATCH (a)-[r2:r2]-()
with a, r1, collect(distinct r2) as r2
....
OPTIONAL MATCH (a)-[problem:r7]-()
RETURN a, r1, r2, r3, r4, r5, r6, collect(distinct propblem) as r7
On Tue, Oct 28, 2014 at 1:55 AM, Mark Angrish <[email protected]>
wrote:
>
> Ok a couple more hours of tinkering and I worked it out. Here is what i
> did in case any one else wants to follow this path of returning graphs back
> rather than rows:
>
> MATCH (a:A{email:"[email protected]"})-[rels:r1|r2|r3|r4|r5|r6|r7]-()
> RETURN user, rels
>
> This effectively returns you all "optional relationships" that you are
> interested in. Returns in under around 100ms.
>
> For what it's worth you can then structure more complexity over the top.
> For me i was interested in bringing down more nodes if a relationship
> condition was met.
>
> MATCH (a:A{email:"[email protected]"})-[rels:r1|r2|r3|r4|r5]-(n)
> WITH a, rels, n
> OPTIONAL MATCH (n:B)-[bRel:r8]-()
> OPTIONAL MATCH (n:C)-[cRel:r9]-()
> RETURN a, rels, bRel, cRel
>
> ::mark
>
>
> On Tuesday, October 28, 2014 9:47:33 AM UTC+11, Mark Angrish wrote:
>>
>> Hi there,
>>
>> I have been trying to solve this bizarre problem for a day now and I
>> can't figure it out. I have the following cypher:
>>
>>
>> MATCH (a:A{email:"[email protected]"})
>> OPTIONAL MATCH (a)-[r1:r1]-()
>> OPTIONAL MATCH (a)-[r2:r2]-()
>> OPTIONAL MATCH (a)-[r3:r3]-()
>> OPTIONAL MATCH (a)-[r4:r4]-()
>> OPTIONAL MATCH (a)-[r5:r5]-()
>> OPTIONAL MATCH (a)-[r6:r6]-()
>> OPTIONAL MATCH (a)-[problem:r7]-()
>> RETURN a, r1, r2, r3, r4, r5, r6, r7
>>
>>
>> There is a unique constraint on A:email. All the other nodes have unique
>> constraints too (usually on a uuid).
>>
>> So I am using the REST API and passing in:
>>
>> "resultDataContents":["graph"]
>>
>> This is to minimise the data sent over the wire and because I map the
>> results to my domain objects. row based/jdbc sends way too much data and
>> can't return properties on relationships AFAIK.
>>
>> The problem i'm having is that by including the relationship problem:r7 my
>> cypher query goes from around 1sec (without it) to around 40 sec (with it).
>> The total data it needs to return is about 100-150 nodes.
>>
>> I have ordered the relationships in order of reducing cardinality to make
>> searching easier. If i do a count distinct return query for
>> a:A-[problem:r7]-() the query returns in about 120ms.
>>
>> Does anyone have any ideas? Also it might matter but it might not: I have
>> upgraded the db several times from 2.1.3 -> 2.1.4 -> 2.1.5.
>>
>>
>>
>> --
> 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/d/optout.
>
--
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/d/optout.
