爲GROUPBY多個實體Neo4j的暗號查詢

Question

Neo4j的暗號查詢來獲取訂單數：GROUPBY一年，然後一個月，然後ORG明智

MATCH (year)-[:MONTH]->(month:Month) 
    WITH year,month 
    MATCH (month)-[r:DAY]->(day:Day) 
    WITH year,month,day 
    OPTIONAL MATCH (order:Order)-[r:ORD_CREATED]->(day) 
    WITH year,month,order 
    MATCH (order)-[:ORDER_OF]-(org) 
    RETURN year,month,COLLECT(DISTINCT(org.name),count(order)) 

我如何能實現以下輸出？任何幫助，將不勝感激

{ 
year 
[{month1,[ {org1,20},{ org2,30}]}], 
[{month2,[ {org1,26},{ org2,10}]}], 
} 

Answer 1

也許不是最優的，但對我的作品：

// match the path from an Org to the Day of order 
MATCH (day:Day)<-[:ORD_CREATED]-(order:Order)-[:ORDER_OF]-(org:Org) 
// count number of paths for each day & org 
WITH day, org, count(*) as opd 
// match on the month for each day 
MATCH (day)<-[:DAY]-(month:Month) 
// sum the order counts for each month & org 
WITH month, org.name as name, SUM(opd) as opm 
// for each month, create a collection of org order counts 
WITH month, COLLECT({ org:name, cnt:opm }) as mon_cnt 
// match on the year for each month 
MATCH (month)<-[:MONTH]-(year:Year) 
// create the final structure of year, list of months, org order counts per month 
RETURN year.year, COLLECT({ month:month.text, mon_cnt:mon_cnt }); 

http://console.neo4j.org/r/y9qyzl的樣品圖和執行結果

|year.year│COLLECT({ month:month.text, mon_cnt:mon_cnt })    │ 
╞═════════╪════════════════════════════════════════════════════════════╡ 
│2015  │[{month: Feb 2015, mon_cnt: [{org: Org5, cnt: 2}, {org: Org2│ 
│   │, cnt: 1}, {org: Org3, cnt: 1}, {org: Org4, cnt: 1}, {org: O│ 
│   │rg1, cnt: 1}]}, {month: Jan 2015, mon_cnt: [{org: Org3, cnt:│ 
│   │ 1}, {org: Org2, cnt: 1}, {org: Org4, cnt: 1}]}]   │ 
├─────────┼────────────────────────────────────────────────────────────┤ 
│2016  │[{month: Feb 2016, mon_cnt: [{org: Org2, cnt: 1}, {org: Org1│ 
│   │, cnt: 1}, {org: Org5, cnt: 1}]}]       │