MongoDB中的explain和hint提的使用
一、简介
mysql适合结构化数据,类似excel表格一样定义严格的数据,用于数据量中,速度一般支持事务处理场合
redis适合缓存内存对象,如缓存队列,用于数据量小,速度快不支持事务处理高并发场合
mongodb,适合半结构化数据,如文本信息,用于数据量大,速度较快不支持事务处理场合
hadoop是个生态系统,上面有大数据分析很多组件,适合事后大数据分析任务
spark类似hadoop,偏向于内存计算,流计算,适合实时半实时大数据分析任务
移动互联网及物联网让数据呈指数增长,NoSql大数据新起后,数据存储领域发展很快,似乎方向都是向大数据,内存计算,分布式框架,平台化发展,出现不少新的方法,普通应用TB,GB级别达不到PB级别的数据存储,用mongodb,mysql就够了,hadoop,spark这类是航母一般多是大规模应用场景,多用于事后分析统计用,如电商的推荐系统分析系统。IAO
看标题,这里是不是跑题了呢,显然不是,了解一下mongodb在存储中的位置还是非常有必要的,explain 和 hint 一看就知道是从mysql借鉴过来的(猜的),实际就是检测查询语句的性能和使用强制索引
二、explain
先写入测试数据
db.test.insertMany([
{ "_id" : 1, "a" : "f1", b: "food", c: 500 },
{ "_id" : 2, "a" : "f2", b: "food", c: 100 },
{ "_id" : 3, "a" : "p1", b: "paper", c: 200 },
{ "_id" : 4, "a" : "p2", b: "paper", c: 150 },
{ "_id" : 5, "a" : "f3", b: "food", c: 300 },
{ "_id" : 6, "a" : "t1", b: "toys", c: 500 },
{ "_id" : 7, "a" : "a1", b: "apparel", c: 250 },
{ "_id" : 8, "a" : "a2", b: "apparel", c: 400 },
{ "_id" : 9, "a" : "t2", b: "toys", c: 50 },
{ "_id" : 10, "a" : "f4", b: "food", c: 75 }]);
写入成功返回值
{
"acknowledged" : true,
"insertedIds" : [
1,
2,
3,
4,
5,
6,
7,
8,
9,
10
]
}
开始查询
> db.test.find();
{ "_id" : 1, "a" : "f1", "b" : "food", "c" : 500 }
{ "_id" : 2, "a" : "f2", "b" : "food", "c" : 100 }
{ "_id" : 3, "a" : "p1", "b" : "paper", "c" : 200 }
{ "_id" : 4, "a" : "p2", "b" : "paper", "c" : 150 }
{ "_id" : 5, "a" : "f3", "b" : "food", "c" : 300 }
{ "_id" : 6, "a" : "t1", "b" : "toys", "c" : 500 }
{ "_id" : 7, "a" : "a1", "b" : "apparel", "c" : 250 }
{ "_id" : 8, "a" : "a2", "b" : "apparel", "c" : 400 }
{ "_id" : 9, "a" : "t2", "b" : "toys", "c" : 50 }
{ "_id" : 10, "a" : "f4", "b" : "food", "c" : 75 }
> db.test.find().count();
10
> db.test.find({ c: { $gte: 100, $lte: 200 }}).count()
3
> db.test.find({ c: { $gte: 100, $lte: 200 }}).explain("executionStats")
{
"queryPlanner" : {
"plannerVersion" : 1,
"namespace" : "test.test",
"indexFilterSet" : false,
"parsedQuery" : {
"$and" : [
{
"c" : {
"$lte" : 200
}
},
{
"c" : {
"$gte" : 100
}
}
]
},
"winningPlan" : {
"stage" : "COLLSCAN",
"filter" : {
"$and" : [
{
"c" : {
"$lte" : 200
}
},
{
"c" : {
"$gte" : 100
}
}
]
},
"direction" : "forward"
},
"rejectedPlans" : [ ]
},
"executionStats" : {
"executionSuccess" : true,
"nReturned" : 3,
"executionTimeMillis" : 0,
"totalKeysExamined" : 0,
"totalDocsExamined" : 10,
"executionStages" : {
"stage" : "COLLSCAN",
"filter" : {
"$and" : [
{
"c" : {
"$lte" : 200
}
},
{
"c" : {
"$gte" : 100
}
}
]
},
"nReturned" : 3,
"executionTimeMillisEstimate" : 0,
"works" : 12,
"advanced" : 3,
"needTime" : 8,
"needYield" : 0,
"saveState" : 0,
"restoreState" : 0,
"isEOF" : 1,
"invalidates" : 0,
"direction" : "forward",
"docsExamined" : 10
}
},
"serverInfo" : {
"host" : "iZbp1g11g0cdeeq9ht9fhjZ",
"port" : 27017,
"version" : "3.4.12",
"gitVersion" : "bfde702b19c1baad532ed183a871c12630c1bbba"
},
"ok" : 1
}
看一下几个关键词
"stage" : "COLLSCAN",
"nReturned" : 3,
"totalDocsExamined" : 10,
全部扫描,不走索引,这里只是演示,所以数据量比较少,如果数据量多起来这样查询将会很慢,甚至会卡死
COLLSCAN
这个是什么意思呢? 如果你仔细一看,应该知道就是CollectionScan,就是所谓的“集合扫描”,对不对,看到集合扫描是不是就可以直接map到数据库中的table scan/heap scan呢??? 是的,这个就是所谓的性能最烂最无奈的由来。
nReturned
这个很简单,就是所谓的numReturned,就是说最后返回的num个数,从图中可以看到,就是最终返回了三条。。。
docsExamined
那这个是什么意思呢??就是documentsExamined,检查了10个documents。。。而从返回上面的nReturned。
创建索引并查询
> db.test.createIndex({ c:1})
{
"createdCollectionAutomatically" : false,
"numIndexesBefore" : 1,
"numIndexesAfter" : 2,
"ok" : 1
}
> db.test.find({ c: { $gte: 100, $lte: 200 }}).explain("executionStats")
{
"queryPlanner" : {
"plannerVersion" : 1,
"namespace" : "test.test",
"indexFilterSet" : false,
"parsedQuery" : {
"$and" : [
{
"c" : {
"$lte" : 200
}
},
{
"c" : {
"$gte" : 100
}
}
]
},
"winningPlan" : {
"stage" : "FETCH",
"inputStage" : {
"stage" : "IXSCAN",
"keyPattern" : {
"c" : 1
},
"indexName" : "c_1",
"isMultiKey" : false,
"isUnique" : false,
"isSparse" : false,
"isPartial" : false,
"indexVersion" : 1,
"direction" : "forward",
"indexBounds" : {
"c" : [
"[100.0, 200.0]"
]
}
}
},
"rejectedPlans" : [ ]
},
"executionStats" : {
"executionSuccess" : true,
"nReturned" : 3,
"executionTimeMillis" : 0,
"totalKeysExamined" : 3,
"totalDocsExamined" : 3,
"executionStages" : {
"stage" : "FETCH",
"nReturned" : 3,
"executionTimeMillisEstimate" : 0,
"works" : 4,
"advanced" : 3,
"needTime" : 0,
"needYield" : 0,
"saveState" : 0,
"restoreState" : 0,
"isEOF" : 1,
"invalidates" : 0,
"docsExamined" : 3,
"alreadyHasObj" : 0,
"inputStage" : {
"stage" : "IXSCAN",
"nReturned" : 3,
"executionTimeMillisEstimate" : 0,
"works" : 4,
"advanced" : 3,
"needTime" : 0,
"needYield" : 0,
"saveState" : 0,
"restoreState" : 0,
"isEOF" : 1,
"invalidates" : 0,
"keyPattern" : {
"c" : 1
},
"indexName" : "c_1",
"isMultiKey" : false,
"isUnique" : false,
"isSparse" : false,
"isPartial" : false,
"indexVersion" : 1,
"direction" : "forward",
"indexBounds" : {
"c" : [
"[100.0, 200.0]"
]
},
"keysExamined" : 3,
"seeks" : 1,
"dupsTested" : 0,
"dupsDropped" : 0,
"seenInvalidated" : 0
}
}
},
"serverInfo" : {
"host" : "iZbp1g11g0cdeeq9ht9fhjZ",
"port" : 27017,
"version" : "3.4.12",
"gitVersion" : "bfde702b19c1baad532ed183a871c12630c1bbba"
},
"ok" : 1
}
再看看上面几个关键词
"stage" : "IXSCAN"
"totalDocsExamined" : 3,
瞬间就少了,这样查询时间也会大大减少
三、hint
这时一个很好玩的一个东西,就是用来force mongodb to excute special index,对吧,为了方便演示,我们做两组复合索引,比如这次我们在c和b上构建一下:
创建索引
> db.test.createIndex({ c:1,b:1})
{
"createdCollectionAutomatically" : false,
"numIndexesBefore" : 2,
"numIndexesAfter" : 3,
"ok" : 1
}
> db.test.createIndex({ b:1,c:1})
{
"createdCollectionAutomatically" : false,
"numIndexesBefore" : 3,
"numIndexesAfter" : 4,
"ok" : 1
}
hint查询
> db.test.find({ c: { $gte: 100, $lte: 200 },b:"food"}).hint({c:1,b:1}).explain("executionStats")
{
"queryPlanner" : {
"plannerVersion" : 1,
"namespace" : "test.test",
"indexFilterSet" : false,
"parsedQuery" : {
"$and" : [
{
"b" : {
"$eq" : "food"
}
},
{
"c" : {
"$lte" : 200
}
},
{
"c" : {
"$gte" : 100
}
}
]
},
"winningPlan" : {
"stage" : "FETCH",
"inputStage" : {
"stage" : "IXSCAN",
"keyPattern" : {
"c" : 1,
"b" : 1
},
"indexName" : "c_1_b_1",
"isMultiKey" : false,
"isUnique" : false,
"isSparse" : false,
"isPartial" : false,
"indexVersion" : 1,
"direction" : "forward",
"indexBounds" : {
"c" : [
"[100.0, 200.0]"
],
"b" : [
"[\"food\", \"food\"]"
]
}
}
},
"rejectedPlans" : [ ]
},
"executionStats" : {
"executionSuccess" : true,
"nReturned" : 1,
"executionTimeMillis" : 0,
"totalKeysExamined" : 3,
"totalDocsExamined" : 1,
"executionStages" : {
"stage" : "FETCH",
"nReturned" : 1,
"executionTimeMillisEstimate" : 10,
"works" : 3,
"advanced" : 1,
"needTime" : 1,
"needYield" : 0,
"saveState" : 0,
"restoreState" : 0,
"isEOF" : 1,
"invalidates" : 0,
"docsExamined" : 1,
"alreadyHasObj" : 0,
"inputStage" : {
"stage" : "IXSCAN",
"nReturned" : 1,
"executionTimeMillisEstimate" : 10,
"works" : 3,
"advanced" : 1,
"needTime" : 1,
"needYield" : 0,
"saveState" : 0,
"restoreState" : 0,
"isEOF" : 1,
"invalidates" : 0,
"keyPattern" : {
"c" : 1,
"b" : 1
},
"indexName" : "c_1_b_1",
"isMultiKey" : false,
"isUnique" : false,
"isSparse" : false,
"isPartial" : false,
"indexVersion" : 1,
"direction" : "forward",
"indexBounds" : {
"c" : [
"[100.0, 200.0]"
],
"b" : [
"[\"food\", \"food\"]"
]
},
"keysExamined" : 3,
"seeks" : 2,
"dupsTested" : 0,
"dupsDropped" : 0,
"seenInvalidated" : 0
}
}
},
"serverInfo" : {
"host" : "iZbp1g11g0cdeeq9ht9fhjZ",
"port" : 27017,
"version" : "3.4.12",
"gitVersion" : "bfde702b19c1baad532ed183a871c12630c1bbba"
},
"ok" : 1
}
正常查询
> db.test.find({ c: { $gte: 100, $lte: 200 },b:"food"}).explain("executionStats")
{
"queryPlanner" : {
"plannerVersion" : 1,
"namespace" : "test.test",
"indexFilterSet" : false,
"parsedQuery" : {
"$and" : [
{
"b" : {
"$eq" : "food"
}
},
{
"c" : {
"$lte" : 200
}
},
{
"c" : {
"$gte" : 100
}
}
]
},
"winningPlan" : {
"stage" : "FETCH",
"inputStage" : {
"stage" : "IXSCAN",
"keyPattern" : {
"b" : 1,
"c" : 1
},
"indexName" : "b_1_c_1",
"isMultiKey" : false,
"isUnique" : false,
"isSparse" : false,
"isPartial" : false,
"indexVersion" : 1,
"direction" : "forward",
"indexBounds" : {
"b" : [
"[\"food\", \"food\"]"
],
"c" : [
"[100.0, 200.0]"
]
}
}
},
"rejectedPlans" : [
{
"stage" : "FETCH",
"inputStage" : {
"stage" : "IXSCAN",
"keyPattern" : {
"c" : 1,
"b" : 1
},
"indexName" : "c_1_b_1",
"isMultiKey" : false,
"isUnique" : false,
"isSparse" : false,
"isPartial" : false,
"indexVersion" : 1,
"direction" : "forward",
"indexBounds" : {
"c" : [
"[100.0, 200.0]"
],
"b" : [
"[\"food\", \"food\"]"
]
}
}
},
{
"stage" : "FETCH",
"filter" : {
"b" : {
"$eq" : "food"
}
},
"inputStage" : {
"stage" : "IXSCAN",
"keyPattern" : {
"c" : 1
},
"indexName" : "c_1",
"isMultiKey" : false,
"isUnique" : false,
"isSparse" : false,
"isPartial" : false,
"indexVersion" : 1,
"direction" : "forward",
"indexBounds" : {
"c" : [
"[100.0, 200.0]"
]
}
}
}
]
},
"executionStats" : {
"executionSuccess" : true,
"nReturned" : 1,
"executionTimeMillis" : 0,
"totalKeysExamined" : 1,
"totalDocsExamined" : 1,
"executionStages" : {
"stage" : "FETCH",
"nReturned" : 1,
"executionTimeMillisEstimate" : 0,
"works" : 3,
"advanced" : 1,
"needTime" : 0,
"needYield" : 0,
"saveState" : 0,
"restoreState" : 0,
"isEOF" : 1,
"invalidates" : 0,
"docsExamined" : 1,
"alreadyHasObj" : 0,
"inputStage" : {
"stage" : "IXSCAN",
"nReturned" : 1,
"executionTimeMillisEstimate" : 0,
"works" : 2,
"advanced" : 1,
"needTime" : 0,
"needYield" : 0,
"saveState" : 0,
"restoreState" : 0,
"isEOF" : 1,
"invalidates" : 0,
"keyPattern" : {
"b" : 1,
"c" : 1
},
"indexName" : "b_1_c_1",
"isMultiKey" : false,
"isUnique" : false,
"isSparse" : false,
"isPartial" : false,
"indexVersion" : 1,
"direction" : "forward",
"indexBounds" : {
"b" : [
"[\"food\", \"food\"]"
],
"c" : [
"[100.0, 200.0]"
]
},
"keysExamined" : 1,
"seeks" : 1,
"dupsTested" : 0,
"dupsDropped" : 0,
"seenInvalidated" : 0
}
}
},
"serverInfo" : {
"host" : "iZbp1g11g0cdeeq9ht9fhjZ",
"port" : 27017,
"version" : "3.4.12",
"gitVersion" : "bfde702b19c1baad532ed183a871c12630c1bbba"
},
"ok" : 1
}
主要对比的还是:
"totalKeysExamined" : 3,
"totalDocsExamined" : 1,
和
"totalKeysExamined" : 1,
"totalDocsExamined" : 1,
是不是比较有意思,有时候monogdb并不会,走你想要的索引,当你创建多个联合索引的时候,情况就比较明显了
MongoDB中的explain和hint提的使用的更多相关文章
- mongodb之使用explain和hint性能分析和优化
当你第一眼看到explain和hint的时候,第一个反应就是mysql中所谓的这两个关键词,确实可以看出,这个就是在mysql中借鉴过来的,既然是借鉴 过来的,我想大家都知道这两个关键字的用处,话不多 ...
- MongoDB的学习--explain()和hint()
Explain 从之前的文章中,我们可以知道explain()能够提供大量与查询相关的信息.对于速度比较慢的查询来说,这是最重要的诊断工具之一.通过查看一个查询的explain()输出信息,可以知道查 ...
- 在MongoDB中执行查询、创建索引
1. MongoDB中数据查询的方法 (1)find函数的使用: (2)条件操作符: (3)distinct找出给定键所有不同的值: (4)group分组: (5)游标: (6)存储过程. 文档查找 ...
- 在MongoDB中执行查询与创建索引
实验目的: (1)掌握MongoDB中数据查询的方法: (2)掌握MongoDB中索引及其创建: 实验内容: 一. MongoDB中数据查询的方法: (1)find函数的使用: (2)条件操作符: a ...
- mongodb中的排序和索引快速学习
在mongodb中,排序和索引其实都是十分容易的,先来小结下排序: 1 先插入些数据 db.SortTest.insert( { name : "Denis", age : ...
- MongoDB中聚合工具Aggregate等的介绍与使用
Aggregate是MongoDB提供的众多工具中的比较重要的一个,类似于SQL语句中的GROUP BY.聚合工具可以让开发人员直接使用MongoDB原生的命令操作数据库中的数据,并且按照要求进行聚合 ...
- MongoDB中的聚合操作
根据MongoDB的文档描述,在MongoDB的聚合操作中,有以下五个聚合命令. 其中,count.distinct和group会提供很基本的功能,至于其他的高级聚合功能(sum.average.ma ...
- MongoDB 大数据技术之mongodb中在嵌套子文档的文档上面建立索引
一.给collection objectid赋自定义的值 MongoDB Enterprise > db.testid.insert({_id:{imsi:"4567890123&qu ...
- MongoDB 索引 和 explain 的使用
索引基本使用 索引是对数据库表中一列或多列的值进行排序的一种结构,可以让我们查询数据库变得 更快.MongoDB 的索引几乎与传统的关系型数据库一模一样,这其中也包括一些基本的查 询优化技巧. 首先我 ...
随机推荐
- Django 2.0 Middleware的写法
网上很多写法,都是传统的写法, process_request和process_response方法,还可以用,但process_view的执行流程已经不行了. 看了官方文档,推荐的写法,也是用__c ...
- BZOJ3626 [LNOI2014]LCA 树链剖分 线段树
欢迎访问~原文出处——博客园-zhouzhendong 去博客园看该题解 题目传送门 - BZOJ3626 题意概括 给出一个n个节点的有根树(编号为0到n-1,根节点为0).一个点的深度定义为这个节 ...
- BZOJ1898 [Zjoi2005]Swamp 沼泽鳄鱼 矩阵
欢迎访问~原文出处——博客园-zhouzhendong 去博客园看该题解 题目传送门 - BZOJ1898 题意概括 有一个无向图. 其中,有许多条鱼在以循环的规律出现,比如循环在1,2,3这些点出现 ...
- X分钟速成Python3
参考博客:Python3 从入门到开车 (与以下代码无关) 源代码下载: learnpython3-cn.py Python是由吉多·范罗苏姆(Guido Van Rossum)在90年代早期设计. ...
- Python String 方法详解
官网文档地址:https://docs.python.org/3/library/stdtypes.html#string-methods 官网 公号:软测小生ruancexiaosheng 文档里的 ...
- TensorFlow 常用的函数
TensorFlow 中维护的集合列表 在一个计算图中,可以通过集合(collection)来管理不同类别的资源.比如通过 tf.add_to_collection 函数可以将资源加入一个或多个集合中 ...
- Android应用开发-数据存储和界面展现(一)
常见布局 相对布局(RelativeLayout) 相对布局下控件默认位置都是左上角(左对齐.顶部对齐父元素),控件之间可以重叠 可以相对于父元素上下左右对齐,相对于父元素水平居中.竖直居中.水平竖直 ...
- NOIp模拟赛 现实(DP 拓扑)
题目来源:by lzz \(Description\) 给定一张有向图,求对于哪些点,删除它和它的所有连边后,图没有环. \(n\leq 5\times10^5,m\leq 10^6\). \(Sol ...
- BZOJ.4530.[BJOI2014]大融合(LCT)
题目链接 BZOJ 洛谷 详见这 很明显题目是要求去掉一条边后两边子树sz[]的乘积. LCT维护的是链的信息,那么子树呢? 我们用s_i[x]来记录轻边连向x的子树的和(记作虚儿子),那么sum[x ...
- Python解释数学系列——分位数Quantile
跳转到我的博客 1. 分位数计算案例与Python代码 案例1 Ex1: Given a data = [6, 47, 49, 15, 42, 41, 7, 39, 43, 40, 36],求Q1, ...