如何在PouchDB上模拟聚合函数avg、sum、max、min、count？

Question

有谁知道如何在 PouchDB 数据库上创建聚合函数，例如 avg、sum、max 和 min。我创建了一个简单的应用程序来测试 PouchDB。我仍然不知道如何运行这些命令。提前致谢。

例如。如何获得 "number" 字段的最高、最低或平均值？

My main Ionic 2 component

import {Component} from '@angular/core';
import {Platform, ionicBootstrap} from 'ionic-angular';
import {StatusBar} from 'ionic-native';
import {HomePage} from './pages/home/home';
declare var require: any;
var pouch = require('pouchdb');
var pouchFind = require('pouchdb-find');
@Component({
    template: '<ion-nav [root]="rootPage"></ion-nav>'
})
export class MyApp {
    rootPage: any = HomePage;
    db: any;
    value: any;
    constructor(platform: Platform) {
        platform.ready().then(() => {
            StatusBar.styleDefault();
        });
        pouch.plugin(pouchFind);
        this.db = new pouch('friendsdb');
        let docs = [
            {
                '_id': '1',
                'number': 10,
                'values': '1, 2, 3',
                'loto': 'fooloto'
            },
            {
                '_id': '2',
                'number': 12,
                'values': '4, 7, 9',
                'loto': 'barloto'
            },
            {
                '_id': '3',
                'number': 13,
                'values': '9, 4, 5',
                'loto': 'fooloto'
            }
        ];
        this.db.bulkDocs(docs).then(function (result) {
            console.log(result);
        }).catch(function (err) {
            console.log(err);
        });
    }
}
ionicBootstrap(MyApp);

Answer 1

您可以使用 map/reduce 函数 of the db.query() method from PouchDB 获取文档的平均值、总和、最大值或任何其他类型的聚合。

我创建了一个 demo JSBin fiddle with a running example。我将功能的解释直接添加到代码（下面）中作为注释，因为我认为这样会更简单。

var db = new PouchDB('friendsdb');
var docs = [
      {'_id': '1', 'number': 10, 'values': '1, 2, 3', 'loto': 'fooloto'},
      {'_id': '2', 'number': 12, 'values': '4, 7, 9', 'loto': 'barloto'},
      {'_id': '3', 'number': 13, 'values': '9, 4, 5', 'loto': 'fooloto'}
];

db.bulkDocs(docs).then(function(result) {
  querySum();
  queryLargest();
  querySmallest();
  queryAverage();
}).catch(function(err) {
  console.log(err);
});

function querySum() {
  function map(doc) {
    // the function emit(key, value) takes two arguments
    // the key (first) arguments will be sent as an array to the reduce() function as KEYS
    // the value (second) arguments will be sent as an array to the reduce() function as VALUES
    emit(doc._id, doc.number);
  }
  function reduce(keys, values, rereduce) {
    // keys:
    //   here the keys arg will be an array containing everything that was emitted as key in the map function...
    //   ...plus the ID of each doc (that is included automatically by PouchDB/CouchDB).
    //   So each element of the keys array will be an array of [keySentToTheEmitFunction, _idOfTheDoc]
    //
    // values
    //   will be an array of the values emitted as value
    console.info('keys ', JSON.stringify(keys));
    console.info('values ', JSON.stringify(values));
    // check for more info: http://couchdb.readthedocs.io/en/latest/couchapp/views/intro.html


    // So, since we want the sum, we can just sum all items of the values array
    // (there are several ways to sum an array, I'm just using vanilla for to keep it simple)
    var i = 0, totalSum = 0;
    for(; i < values.length; i++){
        totalSum += values[i];
    }
    return totalSum;
  }
  db.query({map: map, reduce: reduce}, function(err, response) {
    console.log('sum is ' + response.rows[0].value);
  });
}

function queryLargest() {
  function map(doc) {
    emit(doc._id, doc.number);
  }
  function reduce(keys, values, rereduce) {
    // everything same as before (see querySum() above)
    // so, this time we want the larger element of the values array

    // 
    return Math.max.apply(Math, values);
  }
  db.query({map: map, reduce: reduce}, function(err, response) {
    console.log('largest is ' + response.rows[0].value);
  });
}

function querySmallest() {
  function map(doc) {
    emit(doc._id, doc.number);
  }
  function reduce(keys, values, rereduce) {
    // all the same... now the looking for the min
    return Math.min.apply(Math, values);
  }
  db.query({map: map, reduce: reduce}, function(err, response) {
    console.log('smallest is ' + response.rows[0].value);
  });
}

function queryAverage() {
  function map(doc) {
    emit(doc._id, doc.number);
  }
  function reduce(keys, values, rereduce) {
    // now simply calculating the average
    var i = 0, totalSum = 0;
    for(; i < values.length; i++){
        totalSum += values[i];
    }
    return totalSum/values.length;
  }
  db.query({map: map, reduce: reduce}, function(err, response) {
    console.log('average is ' + response.rows[0].value);
  });
}

注意：这只是一种方法。还有其他几种可能性（不发出 ID 作为键，使用组和不同的 reduce 函数，使用内置的 reduce 函数，例如 _sum，...），我只是认为一般来说这是更简单的选择。

Answer 2

对于此类问题，我是 PouchDB 中 views 的粉丝。

https://pouchdb.com/2014/05/01/secondary-indexes-have-landed-in-pouchdb.html

可以创建一个允许您多次重新查询同一索引的存储视图：这意味着虽然第一次通过很慢（全扫描），但以后的查询会快得多，因为数据已经被索引.

var db = new PouchDB('friendsdb');

var view = {
  '_id':'_design/metrics',
  'views':{
    'metrics':{
      'map':function(doc){
        // collect up all the data we are looking for
        emit(doc._id, doc.number);
      }.toString(),
      'reduce':function(keys, values, rereduce){
        var metrics = {
          sum:0,
          max:Number.MIN_VALUE,
          min:Number.MAX_VALUE,
          count:0
        };
        // aggregate up the values
        for(var i=values.length-1; i>=0; i--){
          var v = values[i];
          metrics.sum += v;
          metrics.max = (metrics.max < v) ? v : metrics.max;
          metrics.min = (metrics.min < v) ? metrics.min : v;
          metrics.count += v.count || 1;
        }
        metrics.avg = metrics.sum/metrics.count;
        return metrics;
      }.toString()
    }
  }
};

// alternately you could use a built in reduce
// if one already exists for the aggregation 
// you are looking for
//view.reduce = '_stats';

// note the addition of the view
var docs = [view
  ,{'_id':'1','number':10,'values':[1,2,3],'loto':'fooloto'}
  ,{'_id':'2','number':12,'values':[4,7,9],'loto':'barloto'}
  ,{'_id':'3','number':13,'values':[9,4,5],'loto':'fooloto'}
];

db.bulkDocs(docs).then(function(result) {
  db.query('metrics',{reduce:true},function(err, response) {
    var m = response.rows[0].value;
    console.log('smallest.: ' + m.min);
    console.log('largest..: ' + m.max);
    console.log('average..: ' + m.avg);
    console.log('count....: ' + m.count);
    console.log('Total ...: ' + m.sum);
  });
}).catch(function(err) {
  console.log(err);
});

请注意将视图添加到您加载到数据库中的数据，以及需要将 map 和 reduce 转换为字符串的事实（函数末尾的 .toString() )

Answer 3

可以使用内置的 _stats reduce 函数检索数字字段的最高值和最低值。

var myMapReduceFun = {
  map: function (doc) {
    emit(doc._id, doc.number);
  },
  reduce: '_stats'
};

db.query(myMapReduceFun, {reduce: true}).then(function (result) {
  // handle result
}).catch(function (err) {
  // handle errors
});

结果类似于：

{"sum":35,"count":3,"min":10,"max":13,"sumsqr":214}

最高值在 "max" 字段中，最低值在 "min" 字段中。现在你只需要计算你想要的平均值，例如平均平均值：

var meanAverage = result.sum / result.count;

PouchDB 中的其他 built-in reduce functions 是 _count 和 _sum。

PouchDB documentation 关于 reduce 函数的说明如下：

Tip: if you’re not using a built-in, you’re probably doing it wrong.

如何在PouchDB上模拟聚合函数avg、sum、max、min、count？

How to simulating the aggregate functions avg, sum, max, min, and count on PouchDB?

javascript

database

nosql

indexeddb

pouchdb