我有一个集合,其内容如下:
[
{
date: "01-01-2001",
hour: 7,
level2: [
{
name: "level2A",
level3: [
{
name: "level3A",
statistic1: 1,
statistic2: 1,
},
{
name: "level3B",
statistic1: 2,
statistic2: 2,
}
]
},
{
name: "level2B",
level3: [
{
name: "level3A",
statistic1: 3,
statistic2: 3,
},
{
name: "level3B",
statistic1: 4,
statistic2: 4,
}
]
}
]
},
{
date: "01-01-2001",
hour: 8,
level2: [
{
name: "level2A",
level3: [
{
name: "level3A",
statistic1: 5,
statistic2: 5,
},
{
name: "level3B",
statistic1: 6,
statistic2: 6,
}
]
},
{
name: "level2B",
level3: [
{
name: "level3A",
statistic1: 7,
statistic2: 7,
},
]
}
]
}
]
在第一级,我有一个日期,一小时和一组level2元素。我每小时都有一个元素。
在第二级,我有一个名称和一个level3元素的集合。请注意,该名称对于当前数组是唯一的,但可以在不同小时的对象中重复。
在最后一级,我有一个名字和一些数字。请注意,该名称对于当前数组是唯一的,但可以在不同level2元素的对象中复制。
我需要按日期汇总数据,统计数字将通过键汇总:date + level2.name + level3.name,而不会破坏复杂对象的结构。对于我在此发布的示例内容,我希望得到以下输入结果:{date:" 01-01-2001"}:
[
{
date: "01-01-2001",
hour: null,
level2: [
{
name: "level2A",
level3: [
{
name: "level3A",
statistic1: 6,
statistic2: 6,
},
{
name: "level3B",
statistic1: 8,
statistic2: 8,
}
]
},
{
name: "level2B",
level3: [
{
name: "level3A",
statistic1: 10,
statistic2: 10,
},
{
name: "level3B",
statistic1: 4,
statistic2: 4,
}
]
}
]
},
]
如何在不破坏结构然后在Node.js代码中重建它的情况下执行此操作?
答案 0 :(得分:1)
尝试以下内容。首先,展开所有嵌套数组,组和组;总和3级统计字段然后组&按3级,然后组&推动2级,最后将字段投影到结果中。
db.collection.aggregate([{
$unwind: "$level2"
}, {
$unwind: "$level2.level3"
}, {
$group: {
_id: {
date: "$date",
level2: "$level2.name",
level3: "$level2.level3.name"
},
statistic1: {
$sum: "$level2.level3.statistic1"
},
statistic2: {
$sum: "$level2.level3.statistic2"
}
}
}, {
$group: {
_id: {
date: "$_id.date",
level2: "$_id.level2"
},
level3: {
$push: {
name: "$_id.level3",
statistic1: "$statistic1",
statistic2: "$statistic2"
}
}
}
}, {
$group: {
_id: "$_id.date",
level2: {
$push: {
name: "$_id.level2",
level3: "$level3"
}
}
}
}, {
$project: {
_id: 0,
date: "$_id",
level2: "$level2"
}
}]);
输出:
{
"level2": [{
"name": "level2A",
"level3": [{
"name": "level3B",
"statistic1": 8,
"statistic2": 8
}, {
"name": "level3A",
"statistic1": 6,
"statistic2": 6
}]
}, {
"name": "level2B",
"level3": [{
"name": "level3B",
"statistic1": 4,
"statistic2": 4
}, {
"name": "level3A",
"statistic1": 10,
"statistic2": 10
}]
}],
"date": "01-01-2001"
}