1. 程式人生 > >【Mongodb】檢視 && 索引

【Mongodb】檢視 && 索引

 

準備工作

準備2個集合的資料,後面檢視和索引都會用到
1個訂單集合,一個收款資訊集合

var orders = new Array();
var shipping = new Array();
var addresses = ["廣西省玉林市", "湖南省岳陽市", "湖北省荊州市", "甘肅省蘭州市", "吉林省松原市", "江西省景德鎮", "遼寧省瀋陽市", "福建省廈門市", "廣東省廣州市", "北京市朝陽區"];

for (var i = 10000; i < 20000; i++) {
    var orderNo = i + Math.random().toString().substr(2, 5);
    orders[i] = { orderNo: orderNo, userId: i, price: Math.round(Math.random() * 10000) / 100, qty: Math.floor(Math.random() * 10) + 1, orderTime: new Date(new Date().setSeconds(Math.floor(Math.random() * 10000))) };

    var address = addresses[Math.floor(Math.random() * 10)];
    shipping[i] = { orderNo: orderNo, address: address, recipienter: "Wilson", province: address.substr(0, 3), city: address.substr(3, 3) }
}
db.order.insert(orders);
db.shipping.insert(shipping);

 

檢視

概述

A MongoDB view is a queryable object whose contents are defined by an aggregation pipeline on other collections or views. MongoDB does not persist the view contents to disk. A view’s content is computed on-demand when a client queries the view. MongoDB can require clients to have permission to query the view. MongoDB does not support write operations against views.

Mongodb的檢視基本上和SQL的檢視一樣

  • 資料來源(集合或檢視)
  • 提供查詢
  • 不實際儲存硬碟
  • 客戶端發起請求查詢時計算而得

1. 建立檢視

有兩種方法建立檢視

db.createCollection(
  "<viewName>",
  {
    "viewOn" : "<source>",
    "pipeline" : [<pipeline>],
    "collation" : { <collation> }
  }
)
db.createView(
  "<viewName>",
  "<source>",
  [<pipeline>],
  {
    "collation" : { <collation> }
  }
)

一般使用db.createView

viewName : 必須,檢視名稱

source : 必須,資料來源,集合/檢視

[<pipeline>] : 可選,一組管道,可見管道是Mongodb比較重要的一環

 

1.1 單個集合建立檢視

假設現在檢視當天最高的10筆訂單檢視,例如後臺某個地方需要實時顯示金額最高的訂單

db.createView(
    "orderInfo",         //檢視名稱
    "order",             //資料來源   
    [
        //篩選符合條件的訂單,大於當天,這裡要注意時區
        { $match: { "orderTime": { $gte: ISODate("2020-04-13T16:00:00.000Z") } } },
        //按金額倒序
        { $sort: { "price": -1 } },
        //限制10個文件
        { $limit: 10 },
        //選擇要顯示的欄位
        //0: 排除欄位,若欄位上使用(_id除外),就不能有其他包含欄位
        //1: 包含欄位
        { $project: { _id: 0, orderNo: 1, price: 1, orderTime: 1 } }
    ]
)

然後就可以直接使用orderInfo這個檢視查詢資料

db.orderInfo.find({})

返回結果

{ "orderNo" : "1755149436", "price" : 100, "orderTime" : ISODate("2020-04-14T13:49:42.220Z") }
{ "orderNo" : "1951423853", "price" : 99.99, "orderTime" : ISODate("2020-04-14T15:08:07.240Z") }
{ "orderNo" : "1196303215", "price" : 99.99, "orderTime" : ISODate("2020-04-14T15:15:41.158Z") }
{ "orderNo" : "1580069456", "price" : 99.98, "orderTime" : ISODate("2020-04-14T13:41:07.199Z") }
{ "orderNo" : "1114480559", "price" : 99.98, "orderTime" : ISODate("2020-04-14T13:31:58.150Z") }
{ "orderNo" : "1229542817", "price" : 99.98, "orderTime" : ISODate("2020-04-14T15:15:35.162Z") }
{ "orderNo" : "1208031402", "price" : 99.94, "orderTime" : ISODate("2020-04-14T14:13:02.160Z") }
{ "orderNo" : "1680622670", "price" : 99.93, "orderTime" : ISODate("2020-04-14T15:17:25.210Z") }
{ "orderNo" : "1549824953", "price" : 99.92, "orderTime" : ISODate("2020-04-14T13:09:41.196Z") }
{ "orderNo" : "1449930147", "price" : 99.92, "orderTime" : ISODate("2020-04-14T15:16:15.187Z") }
 

1.2 多個集合建立檢視

其實跟單個是集合是一樣,只是多了$lookup連線操作符,檢視根據管道最終結果顯示,所以可以關聯多個集合(若出現這種情況就要考慮集合設計是否合理,mongodb本來就是文件型資料庫)

db.orderDetail.drop()
db.createView(
    "orderDetail",
    "order",
    [
        { $lookup: { from: "shipping", localField: "orderNo", foreignField: "orderNo", as: "shipping" } },
        { $project: { "orderNo": 1, "price": 1, "shipping.address": 1 } }
    ]
)

查詢檢視,得到如下結果

{ "_id" : ObjectId("5e95af8c4ef6faf974b4a6c3"), "orderNo" : "1000039782", "price" : 85.94, "shipping" : [ { "address" : "北京市朝陽區" } ] }
{ "_id" : ObjectId("5e95af8c4ef6faf974b4a6c4"), "orderNo" : "1000102128", "price" : 29.04, "shipping" : [ { "address" : "吉林省松原市" } ] }
{ "_id" : ObjectId("5e95af8c4ef6faf974b4a6c5"), "orderNo" : "1000214514", "price" : 90.69, "shipping" : [ { "address" : "湖南省岳陽市" } ] }
{ "_id" : ObjectId("5e95af8c4ef6faf974b4a6c6"), "orderNo" : "1000337987", "price" : 75.05, "shipping" : [ { "address" : "遼寧省瀋陽市" } ] }
{ "_id" : ObjectId("5e95af8c4ef6faf974b4a6c7"), "orderNo" : "1000468969", "price" : 76.84, "shipping" : [ { "address" : "江西省景德鎮" } ] }
{ "_id" : ObjectId("5e95af8c4ef6faf974b4a6c8"), "orderNo" : "1000572219", "price" : 60.25, "shipping" : [ { "address" : "江西省景德鎮" } ] }
{ "_id" : ObjectId("5e95af8c4ef6faf974b4a6c9"), "orderNo" : "1000611743", "price" : 19.14, "shipping" : [ { "address" : "廣東省廣州市" } ] }
{ "_id" : ObjectId("5e95af8c4ef6faf974b4a6ca"), "orderNo" : "1000773917", "price" : 31.5, "shipping" : [ { "address" : "北京市朝陽區" } ] }
{ "_id" : ObjectId("5e95af8c4ef6faf974b4a6cb"), "orderNo" : "1000879146", "price" : 76.16, "shipping" : [ { "address" : "吉林省松原市" } ] }
{ "_id" : ObjectId("5e95af8c4ef6faf974b4a6cc"), "orderNo" : "1000945977", "price" : 93.98, "shipping" : [ { "address" : "遼寧省瀋陽市" } ] }

可以看到,mongodb不是像SQL那樣把連線的表當成列列出,而是把連線結果放在數組裡面,這很符合Mongodb文件型結構。

 

2. 修改檢視

假設現在需要增加一個數量的欄位

db.runCommand({
    collMod: "orderInfo",
    viewOn: "order",
    pipeline: [
        { $match: { "orderTime": { $gte: ISODate("2020-04-13T16:00:00.000Z") } } },
        { $sort: { "price": -1 } },
        { $limit: 10 },
        //增加qty
        { $project: { _id: 0, orderNo: 1, price: 1, qty: 1, orderTime: 1 } }
    ]
})

當然,也可以刪除檢視,重新用db.createView()建立檢視

 

3. 刪除檢視

db.orderInfo.drop();

 

索引

概述

Indexes support the efficient execution of queries in MongoDB. Without indexes, MongoDB must perform a collection scan, i.e. scan every document in a collection, to select those documents that match the query statement. If an appropriate index exists for a query, MongoDB can use the index to limit the number of documents it must inspect.

索引能提供高效的查詢,沒有索引的查詢,mongole執行集合掃描,相當於SQL SERVER的全表掃描,掃描每一個文件。

資料存在儲存介質上,大多數情況是為了查詢,查詢的快慢直接影響使用者體驗,mongodb索引也是空間換時間,新增索引,CUD操作都會導致索引重新生成,影響速度。

 

1. 準備工作

1.1 準備200W條資料

var orderNo = 100 * 10000;
for (var i = 0; i < 100; i++) {
    //分批次插入,每次20000條
    var orders = new Array();
    for (var j = 0; j < 20000; j++) {
        var orderNo = orderNo++;
        orders[j] = { orderNo: orderNo, userId: i + j, price: Math.round(Math.random() * 10000) / 100, qty: Math.floor(Math.random() * 10) + 1, orderTime: new Date(new Date().setSeconds(Math.floor(Math.random() * 10000))) };
    }
    //不需寫入確認
    db.order.insert(orders, { writeConcern: { w: 0 } });
}

 

1.2 mongodb的查詢計劃

db.collection.explain().<method(...)>

 

一般使用執行統計模式,例如

db.order.explain("executionStats").find({orderNo:1000000})

返回的executionStats物件欄位說明

部分欄位說明

欄位說明
executionSuccess 是否執行成功
nReturned 返回匹配文件數量
executionTimeMillis 執行時間,單位:毫秒
totalKeysExamined 索引檢索數目
totalDocsExamined 文件檢索數目

檢視未加索引前查詢計劃

db.order.explain("executionStats").find({orderNo:1000000})

擷取部分返回結果,可以看出

  • executionTimeMillis : 用時1437毫秒
  • totalDocsExamined : 掃描文件200W
  • executionStages.stage : 集合掃描
"executionStats" : {
    "executionSuccess" : true,
    "nReturned" : 1,
    "executionTimeMillis" : 1437,
    "totalKeysExamined" : 0,
    "totalDocsExamined" : 2000000,
    "executionStages" : {
            "stage" : "COLLSCAN",

 

1.3 檢視當前集合統計資訊

db.order.stats()

擷取部分資訊,可以看出現在儲存檔案大小大概為72M

{
        "ns" : "mongo.order",
        "size" : 204000000,
        "count" : 2000000,
        "avgObjSize" : 102,
        "storageSize" : 74473472,

 

2. 建立索引

db.order.createIndex({ orderNo: 1 }, { name: "ix_orderNo" })

索引名稱不是必須,若不指定,按 欄位名稱_排序型別組合自動生成,索引名稱一旦建立不能修改,若要修改,只能刪除索引重新生成索引,建議還是建索引的時候就把索引名稱設定好。

 

2.1 執行查詢計劃

db.order.explain("executionStats").find({orderNo:1000000})

擷取部分結果,直觀就可以感覺查詢速度有了質的提升,再看查詢計劃更加驚訝

  • nReturned : 匹配到1個文件
  • executionTimeMillis : 0,呃。。
  • totalKeysExamined : 總共檢索了1個索引
  • totalDocsExamined : 總共檢索了1個文件
  • executionStages.stage : FETCH,根據索引去檢索指定文件,像SQL的Index Seek
 "executionStats" : {
                "executionSuccess" : true,
                "nReturned" : 1,
                "executionTimeMillis" : 0,
                "totalKeysExamined" : 1,
                "totalDocsExamined" : 1,
                "executionStages" : {
                        "stage" : "FETCH"

 

這裡只介紹最簡單的單個欄位索引,mongodb還有很多索引

  • 複合索引(Compound Indexes):對多個欄位做索引
  • 多鍵索引(Multikey Indexes): 一個欄位多個值做索引,通常是陣列
  • 全文索引(Text Indexes): 對文字檢索,可以對欄位設定不同權重
  • 萬用字元索引(Wildcard Indexes):可以將物件的所有/指定的值做索引
  • 更多

 

參考文章


Views — MongoDB Manual

Indexes — MongoDB Manual

轉發請標明出處:https://www.cnblogs.com/WilsonPan/p/12704474.html