Use cypher to get the tree of the specified structure

Here’s the table of contents:

Use Cypher Get the tree of the specified structure

One 、 Links to questions from the community

Neo4j Figure database Chinese community ： How to get the tree of the specified structure ？

But at the same level node I want to remove duplicates as an array , For example, below

But the parent node and relationship will be lost after simple processing , I want each node to be converted to a map object , This object contains the original node , And of the parent node id, The name of the relationship is easy to serialize directly on the client
But I tried many ways , This can only be done for a single column , Multiple columns become Cartesian products

Two 、 Write query to realize data encapsulation

2.1 Create sample data

2.2 Cypher Realization

     Encapsulate data hierarchically 【 Get the tree of the specified structure 】, Each node of each layer in the returned result contains the relationship associated with the node ID、 node ID; If you need to include nodes in the returned results 、 Relationship Attribute and type information , Can be in the corresponding Cypher The query location is encapsulated and returned together .

//  Match the path to be processed 
MATCH p=(c:Class {name:' The movie '})-[r1: example ]->(m:Movie)<-[r2:ACTED_IN]-(:Person) 
	//  Collect two layers of relationships 
    WITH COLLECT(DISTINCT r1) AS hc_r1,COLLECT(DISTINCT r2) AS hc_r2
    //  Parse the first layer NODE, The second floor NODE, The third level NODE
    // NODE Object structure ：node- Current layer node ,f_node Represents the parent node ,rel Represents the relationship of the current node 
    WITH REDUCE(l=[],r IN hc_r1 | l+apoc.map.setPairs({},[["node",ID(STARTNODE(r))],["rel",ID(r)]])) AS hc_n1,
    	REDUCE(l=[],r IN hc_r1 | l+apoc.map.setPairs({},[["f_node",ID(STARTNODE(r))],["node",ID(ENDNODE(r))],["rel",ID(r)]])) AS hc_n2,
        REDUCE(l=[],r IN hc_r2 | l+apoc.map.setPairs({},[["f_node",ID(ENDNODE(r))],["node",ID(STARTNODE(r))],["rel",ID(r)]])) AS hc_n3
    //  After the last treatment , Nodes of each layer 、 The associations and parent nodes are ready , The next step is to `node` Weight removal , And then `f_node` and `rel` Collected in an array 
    //  There may be multiple parent nodes and associations of the current node , So you need to add two after weight removal map Field `f_nodes` and `rels`
    WITH hc_n1,hc_n2,hc_n3
    WITH apoc.map.groupByMulti(hc_n1,'node') AS hc_n1_gp,EXTRACT(map IN hc_n1 | map.node) AS hc_n_l1, 
    	apoc.map.groupByMulti(hc_n2,'node') AS hc_n2_gp,EXTRACT(map IN hc_n2 | map.node) AS hc_n_l2, 
        apoc.map.groupByMulti(hc_n3,'node') AS hc_n3_gp,EXTRACT(map IN hc_n3 | map.node) AS hc_n_l3
    //  Encapsulate the data format of each layer finally returned [{cur_node:node,f_nodes:[2,123],rels:[213,23]}]
    WITH REDUCE(l=[],nd IN hc_n_l1 | l+apoc.map.setPairs({},[["cur_node",nd],["f_nodes",apoc.coll.toSet(FILTER(e IN EXTRACT(mp IN apoc.map.get(hc_n1_gp,TOSTRING(nd),NULL,FALSE) | mp.f_node) WHERE e IS NOT NULL))],["rels",apoc.coll.toSet(FILTER(e IN EXTRACT(mp IN apoc.map.get(hc_n1_gp,TOSTRING(nd),NULL,FALSE) | mp.rel) WHERE e IS NOT NULL))]])) AS hc1,
    	REDUCE(l=[],nd IN hc_n_l2 | l+apoc.map.setPairs({},[["cur_node",nd],["f_nodes",apoc.coll.toSet(FILTER(e IN EXTRACT(mp IN apoc.map.get(hc_n2_gp,TOSTRING(nd),NULL,FALSE) | mp.f_node) WHERE e IS NOT NULL))],["rels",apoc.coll.toSet(FILTER(e IN EXTRACT(mp IN apoc.map.get(hc_n2_gp,TOSTRING(nd),NULL,FALSE) | mp.rel) WHERE e IS NOT NULL))]])) AS hc2,
        REDUCE(l=[],nd IN hc_n_l3 | l+apoc.map.setPairs({},[["cur_node",nd],["f_nodes",apoc.coll.toSet(FILTER(e IN EXTRACT(mp IN apoc.map.get(hc_n3_gp,TOSTRING(nd),NULL,FALSE) | mp.f_node) WHERE e IS NOT NULL))],["rels",apoc.coll.toSet(FILTER(e IN EXTRACT(mp IN apoc.map.get(hc_n3_gp,TOSTRING(nd),NULL,FALSE) | mp.rel) WHERE e IS NOT NULL))]])) AS hc3
//  Use apoc.coll.toSet Weight removal list 
RETURN apoc.coll.toSet(hc1) AS hc1,
		apoc.coll.toSet(hc2) AS hc2,
		apoc.coll.toSet(hc3) AS hc3