Question

我有格式的文件：

{
  a: "some-value",
  b: "some-other-value",
  c: {"subfield1": "val1", "subfield2": "val2", ..}
}

并希望使用MongoDB生成类似的输出：

{
  "some-value": {
    "some-other-value": {"subfield1": "val1", "subfield2": "val2"},
    ..
  }, 
  ..
}

在RethinkDB中，此查询有效：

r.table("some_table").getAll("some-value", "some-value-2",.., index: :some_index).map{ |d|
  r.object(
    d['a'],
      r.object(
        d['b'],
        d['c'].pluck("subfield1", "subfield2")
      )
    )
}.reduce{|left,right| left.merge(right)}.default({})

有没有办法使用Mongo查询实现相同的目标？

请注意，这有两个部分：

将所有结果映射到嵌套哈希
切割存储在字段c

Answer 1

使用Javascript重新整形map函数中的文档：

db.collection.mapReduce(
    function(){
        emit(this._id, {
            [this.a]: {
                [this.b]: (({ subfield1, subfield2 }) => ({ subfield1, subfield2 }))(this.c)
            }
        });
    },
    function(){},
    {out: {inline: 1}}
).results.map(function(doc){return doc.value})

编辑有点罗嗦：

db.collection.mapReduce(
    function(){
        emit(this._id, {
            [this.a]: {
                [this.b]: { subfield1: this.c.subfield1, subfield2: this.c.subfield2}
            }
        });
    },
    function(){},
    {out: {inline: 1}}
).results.map(function(doc){return doc.value})

EDIT 2 使用reduce函数作为单个对象返回：

db.collection.mapReduce(
    function(){
        emit(null, {
            [this.a]: {
                [this.b]: (({ subfield1, subfield2 }) => ({ subfield1, subfield2 }))(this.c)
            }
        });
    },
    function(key, val){
        var res = {};
        val.forEach(function(doc) {
            // You may want to resolve conflicts here, if you have
            // more than 1 document with the same values for a and b.
            // In this code the later overwrites the former
            res = Object.assig(res, doc);
        });
        return res;
    },
    {out: {inline: 1}}
).results[0].value

是否可以在MongoDB中重写此RQL查询？

1 个答案: