在 MongoDB 中分组和计数 [英] Group and count in MongoDB

问题描述

我正在尝试对 mongodb 2.0.1 中每个组的元素数量进行分组和计数，但到目前为止没有成功.

I'm trying to group and count the amount of elements for each group in mongodb 2.0.1, but with no success so far.

我的数据库架构如下所示:

My DB schema looks like :

{
    "_id" : ObjectId("4ece7544853b4b0941000000"),
    "ResultSet" : {
            "Results" : [
                    {
                            "quality" : 87,
                            "state" : "Franche-Comté"
                    }
            ]
    }
}

我一直在尝试各种方法，遵循不同的教程，但每次都是相同的结果:一个唯一的空组......我不明白为什么.

I've been trying all sort of methods, following different tutorials, but it is each time the same result : an only null group... which I don't understand why.

到目前为止我写的最好的查询如下:

The best query I have written so far is the following :

db.extract_2000.group( {
            cond: { "ResultSet.Results.quality": {$exists: true} },
            key: {"ResultSet.Results.state": true}, 
            reduce: function(obj, glob) { glob.total++; glob.quality += obj.ResultSet.Results.quality },
            initial: { total: 0, quality: 0 },
            finalize: function(glob) {glob.avgquality = glob.quality / glob.total}
            })

返回(再次):

[
        {
                "ResultSet.Results.state" : null,
                "total" : 2000,
                "quality" : NaN,
                "avgquality" : NaN
        }
]

我做错了什么?

推荐答案

这根本不会像写的那样工作.关键问题在这里:key: {"ResultSet.Results.state": true}.ResultSet.Results 是一个数组.当您要求 ResultSet.Results.state 时，您是在暗示某种类型的 for 循环在这里完成.group 命令根本无法做到这一点.

This simply won't work as written. The key problem is here: key: {"ResultSet.Results.state": true}. ResultSet.Results is an array. When you ask for ResultSet.Results.state you are implying some type of for loop be done here. The group command is simply not capable of this.

请尝试以下 M/R:

map = function() {
  // Note that we emit once per result
  foreach(var i in ResultSet.Results) {
    key = this.ResultSet.Results[i];
    value = { count: 1, 
      quality: this.ResultSet.Results[i].quality,
      avg_quality: 0
    };

    emit(key, value);
  }
}

reduce = function(key, values) {
  // note that results has same fields as emitted value
  var results = { count: 0, quality: 0, avg_quality: 0 };
  foreach(var i in values){
    results.count += values[i].count;
    results.quality += values[i].quality;
    // ignore avg_quality, we don't use it
  }
  return results;
}

您还必须为平均值编写 finalize.

You will also have to write a finalize for the average.

finalize = function(key, value) {
  if (value.count > 0)
    value.avg_quality = value.quality / value.count;

  return value;
}

这篇关于在 MongoDB 中分组和计数的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！