mongodb - 如何使用组和累加器阶段改进 MongoDB 聚合？

Question

查询如下:

将选定的文档与匹配项匹配

按月中的某天分组，然后使用累加器将 URL 和计数添加到返回的数据集中

问题是查询非常复杂，随着数据量的增长，这似乎真的不是很好的性能。在 mongodb 中是否有更简单的方法来实现相同的目标？
输出形状如下所示:

{
  "results": [
    {
      "_id": {
        "month": 2,
        "day": 2,
        "year": 2021
      },
      "urls": [
        {
          "url": "https://shop.mydomain.com/product/golden-axe",
          "count": 20
        },
        {
          "url": "https://shop.mydomain.com/product/phantasy-star",
          "count": 218
        },
        {
          "url": "https://shop.mydomain.com/product/sega-arcades-retro",
          "count": 30
        }
      ],
      "count": 268
    },
    {
      "_id": {
        "month": 2,
        "day": 3,
        "year": 2021
      },
      "urls": [
        {
          "url": "https://shop.mydomain.com/product/golden-axe",
          "count": 109
        },
        {
          "url": "https://shop.mydomain.com/product/phantasy-star",
          "count": 416
        },
        {
          "url": "https://shop.mydomain.com/product/sega-arcades-retro",
          "count": 109
        }
      ],
      "count": 634
    },

        const aggregate = [
      {
        $match: {
          source: 'itemLineView',
          createdAt: {
            $gte: new Date(query.dateGT),
            $lte: new Date(query.dateLT)
          },
          url: { $regex: `^${query.url}` },
        }
      },
      {
        $group: {
          _id: {
            month: {
              $month: '$createdAt'
            },
            day: {
              $dayOfMonth: '$createdAt'
            },
            year: {
              $year: '$createdAt'
            }
          },
          urls: {
            // https://docs.mongodb.com/manual/reference/operator/aggregation/accumulator/#grp._S_accumulator
            $accumulator: {
              init: function (): AccumulatorSourceStats {
                return {
                  origins: []
                };
              },
              // initArgs: [], // Argument arr to pass to the init function
              accumulate: function (state: AccumulatorSourceStats, url: string) {
                const index = state.origins.findIndex(function (origin) {
                  return origin.url === url;
                });
                if (index === -1) {
                  state.origins.push({
                    url: url,
                    count: 1
                  });
                } else {
                  ++state.origins[index].count;
                }
                return state;
              },

              accumulateArgs: ['$url'], // Argument(s) passed to the accumulate function

              merge: function (state1: AccumulatorSourceStats, state2: AccumulatorSourceStats) {
                return {
                  origins: state1.origins.concat(state2.origins)
                };
              },

              finalize: function (state: AccumulatorSourceStats) { // Adjust the state to only return field we need
                const sortByUrl = function (a: AccumulatorSourceStatsOrigin, b: AccumulatorSourceStatsOrigin) {
                  if (a.url < b.url) {
                    return -1;
                  }
                  if (a.url > b.url) {
                    return 1;
                  }
                  return 0;
                };
                return state.origins.sort(sortByUrl);
              },

              lang: 'js'
            }
          },
          count: { $sum: 1 }
        }
      },
      { $sort: { _id: 1 } }
    ];
    return this.model.aggregate(aggregate);

Answer 1

来自 docs :

Executing JavaScript inside of an aggregation operator may decrease performance. Only use the $accumulator operator if the provided pipeline operators cannot fulfill your application’s needs.

避免在 Mongo 管道中使用 javascript 代码被认为是一种很好的做法，这应该仅用作最后的手段。在这种情况下，我们可以通过 $group 避免使用它ing 两次，每个 url 每天一次，然后每天一次。像这样:

db.collection.aggregate([
  {
    $match: {
      source: "itemLineView",
      createdAt: {
        $gte: new Date(query.dateGT),
        $lte: new Date(query.dateLT)
      },
      url: {
        $regex: `^${query.url}`
      },
      
    }
  },
  {
    $group: {
      _id: {
        month: {
          $month: "$createdAt"
        },
        day: {
          $dayOfMonth: "$createdAt"
        },
        year: {
          $year: "$createdAt"
        },
        url: "$url"
      },
      count: {
        $sum: 1
      }
    }
  },
  {
    $group: {
      _id: {
        month: "$_id.month",
        day: "$_id.day",
        year: "$_id.year",
        
      },
      urls: {
        $push: {
          url: "$_id.url",
          count: "$count"
        }
      },
      count: {
        $sum: "$count"
      }
    }
  }
])

Mongo Playground