Database Manual / Time Series / Query

Aggregation and Operator Considerations聚合和运算符考虑因素

Some aggregation stages and operators require special considerations when you use them with time series collections.当您将某些聚合阶段和运算符与时间序列集合一起使用时，需要特别注意。

$geonear

Time series collections only support the $geoNear aggregation stage for sorting geospatial data from queries against 2dsphere indexes. 时间序列集合仅支持$geoNear聚合阶段，用于对2dsphere索引查询中的地理空间数据进行排序。You can't use the $near and $nearSphere operators on time series collections.您不能在时间序列集合上使用$near和$nearSphere运算符。

$merge

You cannot use the $merge aggregation stage to add data from another collection to a time series collection.您不能使用$merge聚合阶段将其他集合中的数据添加到时间序列集合中。

$out

Starting in MongoDB 7.0, you can use the $out aggregation stage to write documents to a time series collection. 从MongoDB 7.0开始，您可以使用$out聚合阶段将文档写入时间序列集合。For more information, see Migrate Data into a Time Series Collection.有关详细信息，请参阅将数据迁移到时间序列集合中。

Frequently Used Operations常用操作

The following aggregation pipeline operators and stages are often used to analyze time series data:以下聚合管道运算符和阶段通常用于分析时间序列数据：

  • $dateAdd: Adds a specified amount of time to a Date object.：向Date对象添加指定的时间量。
  • $dateDiff: Returns the time difference between two dates.：返回两个日期之间的时间差。
  • $dateTrunc: Returns a date that has been truncated to the specific unit.：返回已截断为特定单位的日期。
  • $setWindowFields: Runs calculations on documents in a given window.：对给定窗口中的文档运行计算。

Examples示例

Calculate Average Price per Month计算每月平均价格

Consider a dowJonesTickerData collection that contains documents with the following structure:考虑一个dowJonesTickerData集合，其中包含具有以下结构的文档：

{
   date: ISODate("2020-01-03T05:00:00.000Z"),
   symbol: 'AAPL',
   volume: 146322800,
   open: 74.287498,
   adjClose: 73.486023,
   high: 75.144997,
   low: 74.125,
   close: 74.357498
}

This aggregation pipeline performs the following actions:此聚合管道执行以下操作：

  • Uses $dateTrunc to truncate each document's date to the appropriate month.使用$dateTrunc将每个文档的date截断到相应的月份。
  • Uses $group to group the documents by month and symbol.使用$group按月份和符号对文档进行分组。
  • Uses $avg to calculate the average price per month.使用$avg计算每月的平均价格。
db.dowJonesTickerData.aggregate( [ {
   $group: {
      _id: {
         firstDayOfMonth: {
            $dateTrunc: {
               date: "$date",
               unit: "month"
            }
         },
         symbol: "$symbol"
      },
      avgMonthClose: {
         $avg: "$close"
      }
   }
} ] )

The pipeline returns a set of documents where each document contains the average closing price per month for a particular stock.管道返回一组文档，其中每个文档都包含特定股票的每月平均收盘价。

{
   _id: {
      firstDayOfMonth: ISODate("2020-06-01T00:00:00.000Z"),
      symbol: 'GOOG'
   },
   avgMonthClose: 1431.0477184545455
},
{
   _id: {
      firstDayOfMonth: ISODate("2021-07-01T00:00:00.000Z"),
      symbol: 'MDB'
   },
   avgMonthClose: 352.7314293333333
},
{
   _id: {
      firstDayOfMonth: ISODate("2021-06-01T00:00:00.000Z"),
      symbol: 'MSFT'
   },
   avgMonthClose: 259.01818086363636
}

Calculate a Rolling Average Over 30 Days计算30天内的滚动平均值

Consider a dowJonesTickerData collection that contains documents with the following structure:考虑一个dowJonesTickerData集合，其中包含具有以下结构的文档：

{
   date: ISODate("2020-01-03T05:00:00.000Z"),
   symbol: 'AAPL',
   volume: 146322800,
   open: 74.287498,
   adjClose: 73.486023,
   high: 75.144997,
   low: 74.125,
   close: 74.357498
}

This aggregation pipeline performs the following operations:此聚合管道执行以下操作：

  • Uses $setWindowFields to specify a window of 30 days.使用$setWindowFields指定30天的窗口。
  • Calculates a rolling average of the closing price over the last 30 days for each stock.计算每只股票过去30天收盘价的滚动平均值。
db.dowJonesTickerData.aggregate( [
   { $setWindowFields: {
      partitionBy: { symbol : "$symbol" } ,
      sortBy: { date: 1 },
      output: {
         averageMonthClosingPrice: {
            $avg : "$close",
            window : { range : [-1, "current"], unit : "month" }

         }
      }
   } }
] )

The pipeline returns a set of documents where each document includes a $averageMonthClosingPrice field that contains the average of the previous month's closing price for that stock symbol.管道返回一组文档，其中每个文档都包含一个$averageMonthClosingPrice字段，该字段包含该股票代码上个月收盘价的平均值。

{
   date: ISODate("2020-01-29T05:00:00.000Z"),
   symbol: 'AAPL',
   volume: 216229200,
   adjClose: 80.014801,
   low: 80.345001,
   high: 81.962502,
   open: 81.112503,
   close: 81.084999,
   averageMonthClosingPrice: 77.63137520000001
}