neo4j Cypher hierarchical tree build response to JSON

问题

Can you help me to build cypher query? i have following graph db structure:

(parent:Category)-[:subcategory]->(child:Category)

With this graph data i have hierarchical tree with deep level.

I found following code on Stackoverfllow.com and changed for my data:

MATCH (root:Category)-[:subcategory]->(parent:Category)-[:subcategory]->(child:Category)
WITH root, {category: parent, children: collect(child)} AS parent_with_children
WHERE NOT(()-[:subcategory]->(root))
RETURN {category: root, children: collect(parent_with_children)}

But he is build response only for depth with 3 levels of tree. I need bigger. I'm try to build json response like this example:

  [
    category: {
      name: "PC"
      children: {
        category: {
          name: "Parts"
          children: {
            category: {
              name: "CPU"
              ...
            }
          }
        },
        category: {
          name: "Accessories"
          ...
        }
      } 
    }, 
    category: {
      name: "Laptop"
      ...
    }
  ]

The Cypher can make recursive calls? I think this will be better.

Thanks.

P.S. I know there are similar questions on SO, but they did not help me.

回答1:

Cypher is not well suited for dumping out graph data in a tree structure when leaves are at arbitrary depths.

However, with neo4j 3.x, you can get close to what you want if you are able to install the APOC plugin on your server and use the apoc.convert.toTree procedure.

First, let's create some sample data:

CREATE
  (c1:Category {name: 'PC'}),
    (c1)-[:subcategory]->(c2:Category {name: 'Parts'}),
      (c2)-[:subcategory]->(c3:Category {name: 'CPU'}),
        (c3)-[:subcategory]->(c4:Category {name: 'CacheRAM'}),
    (c1)-[:subcategory]->(c5:Category {name: 'Accessories'}),
      (c5)-[:subcategory]->(c6:Category {name: 'Mouse'}),
      (c5)-[:subcategory]->(c7:Category {name: 'Keyboard'}),
  (c10:Category {name: 'Laptop'}),
    (c10)-[:subcategory]->(c20:Category {name: 'Parts'}),
      (c20)-[:subcategory]->(c30:Category {name: 'CPU'}),
    (c10)-[:subcategory]->(c40:Category {name: 'Accessories'}),
      (c40)-[:subcategory]->(c50:Category {name: 'Stylus'});

Then with this query:

MATCH p=(n:Category)-[:subcategory*]->(m)
WHERE NOT ()-[:subcategory]->(n)
WITH COLLECT(p) AS ps
CALL apoc.convert.toTree(ps) yield value
RETURN value;

... you will get N result rows, where N is the number of root Category nodes. Here is a snippet of sample results:

{
  ...
      "row": [
        {
          "_id": 150,
          "_type": "Category",
          "name": "PC",
          "subcategory": [
            {
              "_id": 154,
              "_type": "Category",
              "name": "Accessories",
              "subcategory": [
                {
                  "_id": 156,
                  "_type": "Category",
                  "name": "Keyboard"
                },
                {
                  "_id": 155,
                  "_type": "Category",
                  "name": "Mouse"
                }
              ]
            },
            {
              "_id": 151,
              "_type": "Category",
              "name": "Parts",
              "subcategory": [
                {
                  "_id": 152,
                  "_type": "Category",
                  "name": "CPU",
                  "subcategory": [
                    {
                      "_id": 153,
                      "_type": "Category",
                      "name": "CacheRAM"
                    }
                  ]
                }
              ]
            }
          ]
        }
      ],
  ...
      "row": [
        {
          "_id": 157,
          "_type": "Category",
          "name": "Laptop",
          "subcategory": [
            {
              "_id": 158,
              "_type": "Category",
              "name": "Parts",
              "subcategory": [
                {
                  "_id": 159,
                  "_type": "Category",
                  "name": "CPU"
                }
              ]
            },
            {
              "_id": 160,
              "_type": "Category",
              "name": "Accessories",
              "subcategory": [
                {
                  "_id": 161,
                  "_type": "Category",
                  "name": "Stylus"
                }
              ]
            }
          ]
        }
      ],
  ...
}

来源：https://stackoverflow.com/questions/38578114/neo4j-cypher-hierarchical-tree-build-response-to-json