0

我的密码查询：

MATCH p =(o:Order)-[r:seeks*2..8]->(o:Order)
WHERE o.Name="000093" AND ALL(x IN tail(nodes(p)) WHERE SINGLE(y IN    tail(nodes(p)) WHERE x=y))
RETURN extract(n IN nodes(p)| n.Name) AS OrderID, extract(u IN nodes(p)| u.UserName) AS UserName,length(p), endNode(r[0])
ORDER BY length(p)

我想避免在路径中有具有相同属性值的节点，如何避免它们？

["000093","000090","000096","000097","000107","000091","000089","000093"]
["yunis","gio","Anhar","Jhon","**shakilbit**","xalima","**shakilbit**","yunis"]

所以，订单 0000107 和 000089 是由同一个用户名 shakilbit 下达的，有什么办法可以避免将这些订单放在同一路径上，谢谢！NEO4J .. 据我所知，非常有帮助的社区。

4

1 回答 1

0

使用APOC 程序，您可能希望将您的集合作为一个集合（消除重复值）并比较大小。如果存在重复，则集合的大小会更小。

MATCH p =(o:Order)-[r:seeks*2..8]->(o:Order)
WHERE o.Name="000093" AND ALL(x IN tail(nodes(p)) WHERE SINGLE(y IN    tail(nodes(p)) WHERE x=y))
WITH p, o, r, extract(u IN nodes(p)| u.UserName) AS UserName
// need to make some adjustments since first and last nodes are same
WHERE size(UserName) - 1 = size(apoc.coll.toSet(tail(UserName)))
RETURN extract(n IN nodes(p)| n.Name) AS OrderID, UserName, length(p), endNode(r[0])
ORDER BY length(p)

另一种方法是在 UserName 集合上重复您的ALL(x in tail...) WHERE single()...谓词（或者在您现有的 ALL() 谓词中包含此检查，尽管这可能很昂贵）。您可能希望对每个进行 PROFILE 并查看哪个性能更高。

于 2017-02-26T03:10:38.663 回答