MongoDB 3.4 - 获取没有根文档的子文档数组

Question

我有一个projects包含以下文件的集合：

{
        "_id" : ObjectId("589eff3fee3d13019843f55a"),
        "name" : "Project A",
        "desc" : "test",
        "numofvms" : 0,
        "templates" : [
                {
                        "_id" : ObjectId("589e4c14ee3d131bac8b403c")
                },
                {
                        "_id" : ObjectId("589e4c1dee3d131bac8b403d")
                }
        ],
        "nodes" : [
                {
                        "_id" : ObjectId("589eff8f2bb59057c3f9b89d"),
                        "name" : "Node A"
                },
                {
                        "_id" : ObjectId("589eff962bb59057c3f9b89e"),
                        "name" : "Node B"
                },
                {
                        "_id" : ObjectId("589eff982bb59057c3f9b89f"),
                        "name" : "Node C"
                },
                {
                        "_id" : ObjectId("589eff9a2bb59057c3f9b8a0"),
                        "name" : "Node D"
                }
        ],
        "links" : [ ]
}

我尝试接收nodes给定文档（项目）的子文档（字段）数组，但没有根文档。结果应如下所示：

[
        {
                "_id" : ObjectId("589eff8f2bb59057c3f9b89d"),
                "name" : "Node A"
        },
        {
                "_id" : ObjectId("589eff962bb59057c3f9b89e"),
                "name" : "Node B"
        },
        {
                "_id" : ObjectId("589eff982bb59057c3f9b89f"),
                "name" : "Node C"
        },
        {
                "_id" : ObjectId("589eff9a2bb59057c3f9b8a0"),
                "name" : "Node D"
        }
]

我尝试了如下不同的方法，但无法摆脱根文档：

> db.projects.find({_id: ObjectId("589eff3fee3d13019843f55a")}, { "nodes": 1, _id: 0 }).pretty()
{
        "nodes" : [
                {
                        "_id" : ObjectId("589eff8f2bb59057c3f9b89d"),
                        "name" : "Node A"
                },
                {
                        "_id" : ObjectId("589eff962bb59057c3f9b89e"),
                        "name" : "Node B"
                },
                {
                        "_id" : ObjectId("589eff982bb59057c3f9b89f"),
                        "name" : "Node C"
                },
                {
                        "_id" : ObjectId("589eff9a2bb59057c3f9b8a0"),
                        "name" : "Node D"
                }
        ]
}

这可以直接在查询中实现还是我必须在我的应用程序中手动提取数组？

Answer 1

尝试这个：

db.collection.aggregate([
    {
        $unwind: '$nodes'
    },
    {
        $match: {_id: ObjectId("589eff3fee3d13019843f55a") }
    },
    {
        $replaceRoot: { newRoot: "$nodes" }
    }
]).toArray();

你会得到

[
    {
        "_id" : ObjectId("589eff8f2bb59057c3f9b89d"),
        "name" : "Node A"
    },
    {
        "_id" : ObjectId("589eff962bb59057c3f9b89e"),
        "name" : "Node B"
    },
    {
        "_id" : ObjectId("589eff982bb59057c3f9b89f"),
        "name" : "Node C"
    },
    {
        "_id" : ObjectId("589eff9a2bb59057c3f9b8a0"),
        "name" : "Node D"
    }
]

$unwind操作符是nodes从输入文档中解构字段，为每个元素输出一个文档。

然后使用$replaceRoot阶段将nodes文档提升到顶级，丢弃当前顶级字段。

希望这可以帮助。