Neo4j密码查询以找出连接的节点

Question

I have given query like this

WITH ['1000Anthem.txt','1007AW.txt','100Art.txt'] as NDS 
UNWIND RANGE(0, size(NDS)-2) as i 
UNWIND RANGE(i+1, size(NDS)-1) as j 
WITH NDS, NDS[i] as N1, NDS[j] as N2 
MATCH path = (N1)-[*]-(N2) 
WHERE length(path)+1 <=size(NDS) 
 AND ALL(n in nodes(path) WHERE n in NDS) 
RETURN path    

I got the following error

Type mismatch: N1 already defined with conflicting type String (expected Node) (line 2, column 15 (offset: 224)) "MATCH path = (N1)-[*]-(N2) WHERE length(path)+1 <=size(NDS) AND ALL(n in nodes(path) WHERE n in NDS) RETURN path"

Answer 1

Your N1 and N2 variables are bound to strings from your list.

The MATCH afterwards is attempting to use them as nodes, which isn't possible. A string is not a node.

If you want to lookup a node where one of its properties is equal to the string, you'll need a different approach, using different variables for the nodes, and a predicate in the WHERE clause to filter only nodes where the node's property is equal to the string.

EDIT

You haven't provided any context into what these nodes are supposed to be, no labels, and non-descriptive variable names, so I'm just going to make a wild guess and say these are nodes labeled :File, with the property name .

Your lookup and collecting of the nodes at the start of the query would be something like:

WITH ['1000Anthem.txt','1007AW.txt','100Art.txt'] as NDS 
MATCH (f:File)
WHERE f.name in NDS
WITH collect(f) as NDS
...

If you have an index on :File(name), then the index will be used to speed up that lookup. Your NDS variable at this point will be a collection of nodes instead of a collection of strings, so the remaining parts of your query will be syntactically correct.