带有子图聚合的递归查询

Question

I am trying to use Neo4j to write a query that aggregates quantities along a particular sub-graph.

We have two stores Store1 and Store2 one with supplier S1 the other with supplier S2 . We move 100 units from Store1 into Store3 and 200 units from Store2 to Store3 .

We then move 100 units from Store3 to Store4 . So now Store4 has 100 units and approximately 33 originated from supplier S1 and 66 from supplier S2 .

I need the query to effectively return this information, Eg

S1, 33
S2, 66

I have a recursive query to aggregate all the movements along each path

MATCH p=(store1:Store)-[m:MOVE_TO*]->(store2:Store { Name: 'Store4'}) 
RETURN store1.Supplier, reduce(amount = 0, n IN relationships(p) | amount + n.Quantity) AS reduction

Returns:

| store1.Supplier | reduction|
|-------------------- |-------------|
| S1 | 200 |
| S2 | 300 |
| null | 100 |

Desired:

| store1.Supplier | reduction|
|---------------------|-------------|
| S1 | 33.33 |
| S2 | 66.67 |

Answer 1

What about this one :

MATCH (s:Store) WHERE s.name = 'Store4'
MATCH (s)<-[t:MOVE_TO]-()<-[r:MOVE_TO]-(supp)
WITH t.qty as total, collect(r) as movements
WITH total, movements, reduce(totalSupplier = 0, r IN movements | totalSupplier + r.qty) as supCount
UNWIND movements as movement
RETURN startNode(movement).name as supplier, round(100.0*movement.qty/supCount) as pct

Which returns :

supplier    pct
Store1  33
Store2  67
Returned 2 rows in 151 ms

Answer 2

So the following is pretty ugly, but it works for the example you've given.

MATCH (s4:Store { Name:'Store4' })<-[r1:MOVE_TO]-(s3:Store)<-[r2:MOVE_TO*]-(s:Store) 
WITH s3, r1.Quantity as Factor, SUM(REDUCE(amount = 0, r IN r2 | amount + r.Quantity)) AS Total
MATCH (s3)<-[r1:MOVE_TO*]-(s:Store) 
WITH s.Supplier as Supplier, REDUCE(amount = 0, r IN r1 | amount + r.Quantity) AS Quantity, Factor, Total 
RETURN Supplier, Quantity, Total, toFloat(Quantity) / toFloat(Total) * Factor as Proportion

I'm sure it can be improved.