从neo4j rest API 创建json 的正确方法

Question

I want to create the following json

"model":[
    {
        "_id": 123,
        "innerMode1": [
            {
                "_id": 233,
                "_type": "somtype2",
                "innerMode2": [
                    {
                        "_id": 24533,
                        "_type": "somtype3",
                        "innerMode2": [
                            { ....this goes on till depth of 6
                            }
                        ]
                    }
                ]
            }
        ], "_id": 45123,
        "innerMode2": [
            {
                "_id": 23433,
                "_type": "somtype2",
                "innerMode2": [
                    {
                        "_id": 241233,
                        "_type": "somtype3",
                        "innerMode2": [
                            {
                                ....this goes on till depth of 6
                            }
                        ]
                    }
                ]
            }
        ]
        .
        .
        .
        .
        .
        1000 records

cypher something like:

MATCH (c:Capability{SectorID:{varSectorID}})-[r1:CapabilityAdopts_Capability]->(d)-[r2:Capability_Capability]->(e)-[r3:Capability_Capability]->(f)
OPTIONAL MATCH (f)<-[r3:Capability_Capability]-(g)
OPTIONAL MATCH (f)-[r4:KPI_Capability]-(h)
RETURN c,labels(c),r1,Type(r1),d,labels(d),r2,Type(r2),e,labels(e),r3,Type(r3),f,labels(f),r4,Type(r4),g,labels(g),r5,Type(r5),h,labels(h)

CapabilityAdopts_Capability = innerModel1,

Capability_Capability = innerModel2 (direction of arrows are differnt hence data is different)

Capability_Capability = innerMode13 (direction of arrows are differnt hence data is different)

Capability_Capability = innerMode14 (direction of arrows are differnt hence data is different)

---

  var capObjectsContainer= {
          model: []
        };

var mainContainerL2 = jsonData.results[4].data; //this comes from neo after running the above cypher

for (i = 0; i < mainContainerL2.length; i++) {

    var isPartOfCapabilityContainer = {
        isPartOfCapability: []
    };

    var capIsMeasuredByKPIPatternContainer = {
        cpbltyIsMsrdByKpiPttrn: []
    };


    var consistsOfCapContainer = {
        consistsOfCapability: []
    };

    var isconsistsOfCap2Present = false;

    for (j = 0; j < capObjectsContainer.model.length; j++) {

        if (mainContainerL2[i].row[0]["ExternalID"] == capObjectsContainer.model[j]["_id"]) {

            for (k = 0; k < capObjectsContainer.model[j].adoptsCapability.length; k++) {

                if (mainContainerL2[i].row[4]["ExternalID"] == capObjectsContainer.model[j].adoptsCapability[k]["_id"]) {

                    for (l = 0; l < capObjectsContainer.model[j].adoptsCapability[k].consistsOfCapability.length; l++) {

                        if (mainContainerL2[i].row[8]["ExternalID"] == capObjectsContainer.model[j].adoptsCapability[k].consistsOfCapability[l]["_id"]) {

                            for (m = 0; m < capObjectsContainer.model[j].adoptsCapability[k].consistsOfCapability[l].consistsOfCapability.length; m++) {

                                if (mainContainerL2[i].row[12]["ExternalID"] == capObjectsContainer.model[j].adoptsCapability[k].consistsOfCapability[l].consistsOfCapability[m]["_id"]) {

                                    if (mainContainerL2[i].row[20] != null)
                                        AddcapIsMeasuredByKPIPattern(capObjectsContainer.model[j].adoptsCapability[k].consistsOfCapability[l].consistsOfCapability[m].cpbltyIsMsrdByKpiPttrn, mainContainerL2[i].row[20], mainContainerL2[i].row[21])

                                    if (mainContainerL2[i].row[16] != null && !IsExternalIdPresent(capObjectsContainer.model[j].adoptsCapability[k].consistsOfCapability[l].consistsOfCapability[m].isPartOfCapability, mainContainerL2[i].row[16]))
                                        AddIsPartOfCap(capObjectsContainer.model[j].adoptsCapability[k].consistsOfCapability[l].consistsOfCapability[m].isPartOfCapability, mainContainerL2[i].row[16], mainContainerL2[i].row[17])

                                    isconsistsOfCap2Present = true;
                                    break;
                                }
                            }
                            if (!isconsistsOfCap2Present) {
                                if (mainContainerL2[i].row[20] != null)
                                    AddcapIsMeasuredByKPIPattern(capIsMeasuredByKPIPatternContainer.cpbltyIsMsrdByKpiPttrn, mainContainerL2[i].row[20], mainContainerL2[i].row[21])

                                if (mainContainerL2[i].row[16] != null)
                                    AddIsPartOfCap(isPartOfCapabilityContainer.isPartOfCapability, mainContainerL2[i].row[16], mainContainerL2[i].row[17])

                                if (mainContainerL2[i].row[12] != null)
                                    AddConsistsOfCap(capObjectsContainer.model[j].adoptsCapability[k].consistsOfCapability[l].consistsOfCapability, mainContainerL2[i].row[12], mainContainerL2[i].row[13], isPartOfCapabilityContainer.isPartOfCapability, capIsMeasuredByKPIPatternContainer.cpbltyIsMsrdByKpiPttrn, consistsOfCapContainer.consistsOfCapability)

                            }
                            break;
                        }

                    }
                }
            }
        }
    }
}

requirement is there cannot be a round trip to neo. I can hit neo only once and get data. I dont like the above code which i wrote (though it creates json exactly as I want). is there a better cypher query that I can use to accomplish what I want? I want to get rid of all the loops. Any functions that I can use or property that can help me here?

Here is a link .

Answer 1

You can use collect + map and collection syntax to create something like you describe in a single cypher statement, you'd start at the lowest level and then work your way upwards aggregating along your paths.

Unfortunately I was not able to follow your explanation.

Here is a link to a article that explains it in detail:

http://gist.neo4j.org/?9269173