MongoDB聚合平均值（）

嘿，伙计们，我真的是一个新的，如果涉及到聚合，所以请帮助我通过这一点

假设我有多个文档（随时间推移）如下：

{
  "_id": ObjectId("574d6175da461e77030041b7"),
  "hostname": "VPS",
  "timestamp": NumberLong(1460040691),
  "cpuCores": NumberLong(2),
  "cpuList": [
    {
      "name": "cpu1",
      "load": 3.4
    },
    {
      "name": "cpu2",
      "load": 0.7
    }
  ]
},
{
  "_id": ObjectId("574d6175da461e77030041b7"),
  "hostname": "VPS",
  "timestamp": NumberLong(1460040700),
  "cpuCores": NumberLong(2),
  "cpuList": [
    {
      "name": "cpu1",
      "load": 0.4
    },
    {
      "name": "cpu2",
      "load": 6.7
    }
  ]
},
{
  "_id": ObjectId("574d6175da461e77030041b7"),
  "hostname": "VPS",
  "timestamp": NumberLong(1460041000),
  "cpuCores": NumberLong(2),
  "cpuList": [
    {
      "name": "cpu1",
      "load": 25.4
    },
    {
      "name": "cpu2",
      "load": 1.7
    }
  ]
}

{
    "avgCPULoad": "2.8",
    "timestamp": NumberLong(1460040700)
},
{
    "avgCPULoad": "13.55",
    "timestamp": NumberLong(1460041000)
}

db.Pizza.aggregate(
[
    {
        $group: 
        {
            _id:
            {
                $subtract: [
                    '$timestamp',   
                    {
                        $mod: ['$timestamp', 300]
                    }
                ]
            },
            'timestamp': {$last:'$timestamp'}
        },
    {
        $project: {_id: 0, timestamp:'$timestamp'}
    }
])

我想得到X时间内的平均cpu负载。其中X等于300秒

因此，在上面的示例中，我们将得到如下结果集：

{
  "_id": ObjectId("574d6175da461e77030041b7"),
  "hostname": "VPS",
  "timestamp": NumberLong(1460040691),
  "cpuCores": NumberLong(2),
  "cpuList": [
    {
      "name": "cpu1",
      "load": 3.4
    },
    {
      "name": "cpu2",
      "load": 0.7
    }
  ]
},
{
  "_id": ObjectId("574d6175da461e77030041b7"),
  "hostname": "VPS",
  "timestamp": NumberLong(1460040700),
  "cpuCores": NumberLong(2),
  "cpuList": [
    {
      "name": "cpu1",
      "load": 0.4
    },
    {
      "name": "cpu2",
      "load": 6.7
    }
  ]
},
{
  "_id": ObjectId("574d6175da461e77030041b7"),
  "hostname": "VPS",
  "timestamp": NumberLong(1460041000),
  "cpuCores": NumberLong(2),
  "cpuList": [
    {
      "name": "cpu1",
      "load": 25.4
    },
    {
      "name": "cpu2",
      "load": 1.7
    }
  ]
}

{
    "avgCPULoad": "2.8",
    "timestamp": NumberLong(1460040700)
},
{
    "avgCPULoad": "13.55",
    "timestamp": NumberLong(1460041000)
}

db.Pizza.aggregate(
[
    {
        $group: 
        {
            _id:
            {
                $subtract: [
                    '$timestamp',   
                    {
                        $mod: ['$timestamp', 300]
                    }
                ]
            },
            'timestamp': {$last:'$timestamp'}
        },
    {
        $project: {_id: 0, timestamp:'$timestamp'}
    }
])

avgCpuLoad的计算如下：

在相隔300秒内抓取所有文档

计算平均值：

（（3.4+0.7）/2）+（0.4+6.7）/2）/2=2.8

（（25.4+1.7）/2）=13.55

添加所选文档中的最后一个时间戳

我知道我是如何在每次x时间获得每个文档的。就是这样做的：

{
  "_id": ObjectId("574d6175da461e77030041b7"),
  "hostname": "VPS",
  "timestamp": NumberLong(1460040691),
  "cpuCores": NumberLong(2),
  "cpuList": [
    {
      "name": "cpu1",
      "load": 3.4
    },
    {
      "name": "cpu2",
      "load": 0.7
    }
  ]
},
{
  "_id": ObjectId("574d6175da461e77030041b7"),
  "hostname": "VPS",
  "timestamp": NumberLong(1460040700),
  "cpuCores": NumberLong(2),
  "cpuList": [
    {
      "name": "cpu1",
      "load": 0.4
    },
    {
      "name": "cpu2",
      "load": 6.7
    }
  ]
},
{
  "_id": ObjectId("574d6175da461e77030041b7"),
  "hostname": "VPS",
  "timestamp": NumberLong(1460041000),
  "cpuCores": NumberLong(2),
  "cpuList": [
    {
      "name": "cpu1",
      "load": 25.4
    },
    {
      "name": "cpu2",
      "load": 1.7
    }
  ]
}

{
    "avgCPULoad": "2.8",
    "timestamp": NumberLong(1460040700)
},
{
    "avgCPULoad": "13.55",
    "timestamp": NumberLong(1460041000)
}

db.Pizza.aggregate(
[
    {
        $group: 
        {
            _id:
            {
                $subtract: [
                    '$timestamp',   
                    {
                        $mod: ['$timestamp', 300]
                    }
                ]
            },
            'timestamp': {$last:'$timestamp'}
        },
    {
        $project: {_id: 0, timestamp:'$timestamp'}
    }
])

但是怎样才能得到像上面那样的平均值呢？

---

我尝试了一下

$unwind

，但没有给出我想要的结果。

解决方法是在阵列上使用unwind（cpulist）。我为您提供了一个示例查询：

db.CpuInfo.aggregate([
    {
        $unwind: '$cpuList'
    },
    {
        $group: {
            _id:{
                $subtract:[
                    '$timestamp', 
                    {$mod: ['$timestamp', 300]}
                ]
            }, 
            'timestamp':{$last:'$timestamp'},
            'cpuList':{$avg:'$cpuList.load'}
        }
    }
])

---

您需要运行以下聚合操作以获得所需的结果：

db.collection.aggregate([ 
    { "$unwind": "$cpuList" },
    {
        "$group": {
            "_id": {                
                "interval": {
                    "$subtract": [ 
                        "$timestamp",                        
                        { "$mod": [ "$timestamp", 60 * 5 ] }
                    ]
                }
            },             
            "avgCPULoad": { "$avg": "$cpuList.load" },
            "timestamp": { "$max": "$timestamp" }
        } 
    },
    {
        "$project": { "_id": 0, "avgCPULoad": 1, "timestamp": 1 }
    }
])

以上按5分钟的间隔（以秒为单位）对展平的文档进行分组；将实际时间戳除以5分钟间隔（以秒为单位）后得到的余数减去以秒为单位的时间戳，即可得出间隔键

样本输出

/* 1 */
{
    "avgCPULoad" : 13.55,
    "timestamp" : NumberLong(1460041000)
}

/* 2 */
{
    "avgCPULoad" : 2.8,
    "timestamp" : NumberLong(1460040700)
}

---

考虑更改您的模式，以便为您的代码>“CulpSistar”<代码>字段（例如“代码> > CCPIST”）具有嵌入的关键值文档（{名称：“CPU1”，Load：3.4 }，{名称：“CPU2”，加载：0.7 } < /代码>）。您还需要将

加载

值转换为数字值，以便像

$avg

这样的聚合累加器操作符可以有效。我之所以创建加载属性字符串，是因为每当我将其设置为数字值时，我都会对精度浮点造成混乱。例如，在php中，一个浮点数四舍五入为2位小数。每当我把它放在Mongo中，它得到的不是2个小数，比如10多个，或者一个非常小的数字offset@chridam好吧，我已经按照你的建议改变了我的模式。如何进行查询以计算平均值？时间戳以秒为单位，而不是以秒为单位miliseconds@Baklap4你说得对，我的错。我已经更新了答案，以反映秒而不是毫秒。是否有可能在mongo中进行取整？是的，但这会有点麻烦，因为mongo中没有

$round

操作符。您可以使用@Baklap4进行黑客攻击，如果您可以为此创建一个新问题，则最好是这样，以免您遇到失败。