Introduction to NoSql Database MongoDB

Mongo is a high-performance, open source, schema-free document database that can be used to replace traditional relational databases or key-value storage in many scenarios. Mongo is developed in C++ and has the following features:

l Collection-oriented storage: suitable for storing data in the form of objects and JSON .

l Dynamic query: Mongo supports rich query expressions. The query command uses JSON format markup to easily query objects and arrays embedded in the document.

l Complete index support: including document embedded objects and arrays. Mongo's query optimizer analyzes the query expression and generates an efficient query plan.

l Query monitoring: Mongo includes a monitoring tool for analyzing the performance of database operations.

l Replication and automatic failover: Mongo database supports data replication between servers, master-slave mode and mutual replication between servers. The primary goal of replication is to provide redundancy and automatic failover.

l Efficient traditional storage: supports binary data and large objects (such as photos or pictures).

l Automatic sharding to support cloud-level scalability (in early alpha stage): The automatic sharding feature supports horizontal database clusters with the ability to dynamically add additional machines.

Schema-free means that for a file stored in a mongodb database, we do not need to know any structure definition of it. If desired, you can store files of different structures in the same database.

Documents stored in a collection are stored as key-value pairs. The key is used to uniquely identify a document and is of type string, while the value can be of various complex file types. We call this storage form BSON (Binary Serialized dOcument Format).

The MongoDB server can run on Linux, Windows or OS X platforms, supports 32-bit and 64-bit applications, and the default port is 27017. It is recommended to run on a 64-bit platform, as the maximum file size supported by MongoDB when running in 32-bit mode is 2GB.

MongoDB stores data in files (the default path is: /data/db), and uses memory-mapped files for management to improve efficiency.

The main goal of MongoDB is to bridge the gap between key/value storage (which provides high performance and high scalability) and traditional RDBMS systems (rich functionality), combining the best of both worlds.

According to the description on the official website , Mongo is suitable for the following scenarios:

l Website data: Mongo is very suitable for real-time insertion, update and query, and has the replication and high scalability required for real-time website data storage.

l Caching: Due to its high performance, Mongo is also suitable as a caching layer for information infrastructure. After the system is restarted, the persistent cache layer built by Mongo can avoid overloading the underlying data sources.

l Large size, low-value data: It may be expensive to store some data in a traditional relational database. Before that, programmers often chose traditional files for storage.

l High scalability scenarios: Mongo is very suitable for databases consisting of dozens or hundreds of servers. Mongo's roadmap already includes built-in support for the MapReduce engine.

l Storage of objects and JSON data: Mongo's BSON data format is very suitable for storage and query in a documented format.

Naturally, the use of MongoDB will also have some limitations, for example it is not suitable for:

l Highly transactional systems: such as banking or accounting systems. Traditional relational databases are still more suitable for applications that require a large number of atomic complex transactions.

l Traditional business intelligence applications: BI databases for specific problems will generate highly optimized query methods. For such applications, a data warehouse may be a more appropriate choice.

l Questions that require SQL

MongoDB supports OS X, Linux, Windows and other operating systems, and provides drivers for Python, PHP, Ruby, Java and C++ languages. The community also provides drivers for platforms such as Erlang and .NET.

Like mysqld, a mongod service can have multiple databases, each database can have multiple tables, the table here is called collection, each collection can store multiple documents, and each document is BSON (binary) json) is stored on the hard disk. The difference from relational databases is that they are stored in single document units. You can add or delete fields to one or a batch of documents without affecting other documents. This is the so-called schema. -free, which is also the main advantage of document databases. Unlike the general key-value database, its value stores structural information, so you can read, write, and count certain fields like a relational database. It can be said that it has both the convenience and efficiency of the key-value database and the powerful functions of the relational database.

start the database

Download MongoDB , unzip it and start:

$ bin/mongod

By default, MongoDB stores data in /data/db/ (or c:/data/db). Of course, you can change it to a different directory, just specify the --dbpath parameter:

$ bin/mongod --dbpath /path/to/my/data/dir

get database connection

Now we can use the built-in shell tool to operate the database. (We can also use the drivers of various programming languages to use MongoDB, and the built-in shell tool can facilitate us to manage the database)

Start the MongoDB JavaScript tool:

$ bin / mongo

The default shell connects to the test library on the local localhost, and you will see:

MongoDB shell version: 0.9.8
url: test
connecting to: test
type "help" for help
>

“connecting to:” This will display the name of the database you are using. If you want to change the database you can:

> use mydb

You can type help to see all the commands.

insert data into collection

下面我们来建立一个test的集合并写入一些数据. 建立两个对象, j 和 t , 并保存到集合中去.
在例子里 ‘>’ 来表示是 shell 输入提示符

> j = { name : "mongo" };
{"name" : "mongo"}
> t = { x : 3 };
{ "x" : 3  }
> db.things.save(j);
> db.things.save(t);
> db.things.find();
{"name" : "mongo" , "_id" : ObjectId("497cf60751712cf7758fbdbb")}
{"x" : 3 , "_id" : ObjectId("497cf61651712cf7758fbdbc")}
>

有几点需要注意下 :

不需要预先建立一个集合. 在第一次插入数据时候会自动建立.
在例子其实可以存储任何结构的数据, 当然在实际应用我们存储的还是相同元素的集合. 这个特性其实可以在应用里很灵活, 你不需要类似 alter table 来修改你的数据结构
每次插入数据时候对象都会有一个ID, 名字叫 _id.

当你运行不同的例子, 你的对象ID值都是不同的.

下面再加点数据:

> for( var i = 1; i < 10; i++ ) db.things.save( { x:4, j:i } ); > db.things.find();
{"name" : "mongo" , "_id" : ObjectId("497cf60751712cf7758fbdbb")}
{"x" : 3 , "_id" : ObjectId("497cf61651712cf7758fbdbc")}
{"x" : 4 , "j" : 1 , "_id" : ObjectId("497cf87151712cf7758fbdbd")}
{"x" : 4 , "j" : 2 , "_id" : ObjectId("497cf87151712cf7758fbdbe")}
{"x" : 4 , "j" : 3 , "_id" : ObjectId("497cf87151712cf7758fbdbf")}
{"x" : 4 , "j" : 4 , "_id" : ObjectId("497cf87151712cf7758fbdc0")}
{"x" : 4 , "j" : 5 , "_id" : ObjectId("497cf87151712cf7758fbdc1")}
{"x" : 4 , "j" : 6 , "_id" : ObjectId("497cf87151712cf7758fbdc2")}
{"x" : 4 , "j" : 7 , "_id" : ObjectId("497cf87151712cf7758fbdc3")}
{"x" : 4 , "j" : 8 , "_id" : ObjectId("497cf87151712cf7758fbdc4")}

请注意下, 这里循环次数是10, 但是只显示到8, 还有2条数据没有显示.

如果想继续查询下面的数据只需要使用 it 命令, 就会继续下面的数据:

{"x" : 4 , "j" : 7 , "_id" : ObjectId("497cf87151712cf7758fbdc3")}
{"x" : 4 , "j" : 8 , "_id" : ObjectId("497cf87151712cf7758fbdc4")}

继续

> it
{"x" : 4 , "j" : 9 , "_id" : ObjectId("497cf87151712cf7758fbdc5")}
{"x" : 4 , "j" : 10 , "_id" : ObjectId("497cf87151712cf7758fbdc6")}

从技术上讲 find() 返回一个游标对象. 但在上面的例子里, 并没有拿到一个游标的变量. 所以 shell 自动遍历游标, 返回一个初始化的set, 并允许我们继续用 it 迭代输出.

当然我们也可以直接用游标来输出, 不过这个是下一部分的内容了.

查询数据

在没有深入查询之前, 我们先看看怎么从一个查询中返回一个游标对象. 可以简单的通过 find() 来查询, 他返回一个任意结构的集合. 如果实现特定的查询稍后讲解.

实现上面同样的查询, 然后通过 while 来输出:

> var cursor = db.things.find();
> while (cursor.hasNext()) { print(tojson(cursor.next())); }
{"name" : "mongo" , "_id" : ObjectId("497cf60751712cf7758fbdbb")}
{"x" : 3 , "_id" : ObjectId("497cf61651712cf7758fbdbc")}
{"x" : 4 , "j" : 1 , "_id" : ObjectId("497cf87151712cf7758fbdbd")}
{"x" : 4 , "j" : 2 , "_id" : ObjectId("497cf87151712cf7758fbdbe")}
{"x" : 4 , "j" : 3 , "_id" : ObjectId("497cf87151712cf7758fbdbf")}
{"x" : 4 , "j" : 4 , "_id" : ObjectId("497cf87151712cf7758fbdc0")}
{"x" : 4 , "j" : 5 , "_id" : ObjectId("497cf87151712cf7758fbdc1")}
{"x" : 4 , "j" : 6 , "_id" : ObjectId("497cf87151712cf7758fbdc2")}
{"x" : 4 , "j" : 7 , "_id" : ObjectId("497cf87151712cf7758fbdc3")}
{"x" : 4 , "j" : 8 , "_id" : ObjectId("497cf87151712cf7758fbdc4")}
{"x" : 4 , "j" : 9 , "_id" : ObjectId("497cf87151712cf7758fbdc5")}
>

上面的例子显示了游标风格的迭代输出. hasNext() 函数告诉我们是否还有数据, 如果有则可以调用 next() 函数. 这里我们也用了自带的 tojson() 方法返回一个标准的 JSON 格式数据.

当我们使用的是 JavaScript shell, 可以用到JS的特性, forEach 就可以输出游标了. 下面的例子就是使用 forEach() 来循环输出:

> db.things.find().forEach( function(x) { print(tojson(x));});
{"name" : "mongo" , "_id" : ObjectId("497cf60751712cf7758fbdbb")}
{"x" : 3 , "_id" : ObjectId("497cf61651712cf7758fbdbc")}
{"x" : 4 , "j" : 1 , "_id" : ObjectId("497cf87151712cf7758fbdbd")}
{"x" : 4 , "j" : 2 , "_id" : ObjectId("497cf87151712cf7758fbdbe")}
{"x" : 4 , "j" : 3 , "_id" : ObjectId("497cf87151712cf7758fbdbf")}
{"x" : 4 , "j" : 4 , "_id" : ObjectId("497cf87151712cf7758fbdc0")}
{"x" : 4 , "j" : 5 , "_id" : ObjectId("497cf87151712cf7758fbdc1")}
{"x" : 4 , "j" : 6 , "_id" : ObjectId("497cf87151712cf7758fbdc2")}
{"x" : 4 , "j" : 7 , "_id" : ObjectId("497cf87151712cf7758fbdc3")}
{"x" : 4 , "j" : 8 , "_id" : ObjectId("497cf87151712cf7758fbdc4")}
{"x" : 4 , "j" : 9 , "_id" : ObjectId("497cf87151712cf7758fbdc5")}
>

forEach() 必须定义一个函数供每个游标元素调用.

在 mongo shell 里, 我们也可以把游标当作数组来用 :

> var cursor = db.things.find();
> print (tojson(cursor[4]));
{"x" : 4 , "j" : 3 , "_id" : ObjectId("497cf87151712cf7758fbdbf")}

使用游标时候请注意占用内存的问题, 特别是很大的游标对象, 有可能会内存溢出. 所以应该用迭代的方式来输出.

下面的示例则是把游标转换成真实的数组类型:

> var arr = db.things.find().toArray();
> arr[5];
{"x" : 4 , "j" : 4 , "_id" : ObjectId("497cf87151712cf7758fbdc0")}

请注意这些特性只是在 mongo shell 里使用, 而不是所有的其他应用程序驱动都支持.

MongoDB 游标对象不是没有快照 – 如果有其他用户在集合里第一次或者最后一次调用 next(), 你可以得不到游标里的数据. 所以要明确的锁定你要查询的游标.

指定条件的查询

到这里我们已经知道怎么从游标里实现一个查询并返回数据对象, 下面就来看看怎么根据指定的条件来查询.

下面的示例就是说明如何执行一个类似SQL的查询, 并演示了怎么在 MongoDB 里实现. 这是在 MongoDB shell 里查询, 当然你也可以用其他的应用驱动或者语言来实现:

SELECT * FROM things WHERE name="mongo"

> db.things.find({name:"mongo"}).forEach(function(x) { print(tojson(x));});
{"name" : "mongo" , "_id" : ObjectId("497cf60751712cf7758fbdbb")}
>

SELECT * FROM things WHERE x=4

> db.things.find({x:4}).forEach(function(x) { print(tojson(x));});
{"x" : 4 , "j" : 1 , "_id" : ObjectId("497cf87151712cf7758fbdbd")}
{"x" : 4 , "j" : 2 , "_id" : ObjectId("497cf87151712cf7758fbdbe")}
{"x" : 4 , "j" : 3 , "_id" : ObjectId("497cf87151712cf7758fbdbf")}
{"x" : 4 , "j" : 4 , "_id" : ObjectId("497cf87151712cf7758fbdc0")}
{"x" : 4 , "j" : 5 , "_id" : ObjectId("497cf87151712cf7758fbdc1")}
{"x" : 4 , "j" : 6 , "_id" : ObjectId("497cf87151712cf7758fbdc2")}
{"x" : 4 , "j" : 7 , "_id" : ObjectId("497cf87151712cf7758fbdc3")}
{"x" : 4 , "j" : 8 , "_id" : ObjectId("497cf87151712cf7758fbdc4")}
{"x" : 4 , "j" : 9 , "_id" : ObjectId("497cf87151712cf7758fbdc5")}
>

查询条件是 { a:A, b:B, … } 类似 “where a==A and b==B and …”, 更多的查询方式可以参考 Mongo 开发教程部分.

上面显示的是所有的元素, 当然我们也可以返回特定的元素, 类似于返回表里某字段的值, 只需要在 find({x:4}) 里指定元素的名字, 比如 j:

SELECT j FROM things WHERE x=4

> db.things.find({x:4}, {j:true}).forEach(function(x) { print(tojson(x));});
{"j" : 1 , "_id" : ObjectId("497cf87151712cf7758fbdbd")}
{"j" : 2 , "_id" : ObjectId("497cf87151712cf7758fbdbe")}
{"j" : 3 , "_id" : ObjectId("497cf87151712cf7758fbdbf")}
{"j" : 4 , "_id" : ObjectId("497cf87151712cf7758fbdc0")}
{"j" : 5 , "_id" : ObjectId("497cf87151712cf7758fbdc1")}
{"j" : 6 , "_id" : ObjectId("497cf87151712cf7758fbdc2")}
{"j" : 7 , "_id" : ObjectId("497cf87151712cf7758fbdc3")}
{"j" : 8 , "_id" : ObjectId("497cf87151712cf7758fbdc4")}
{"j" : 9 , "_id" : ObjectId("497cf87151712cf7758fbdc5")}
>

请注意 “_id” 元素会一直被返回.

findOne() – 语法糖

为了方便, mongo shell (其他驱动) 避免游标的可能带来的开销, 提供一个findOne() 函数. 这个函数和 find() 参数一样, 不过他返回游标里第一条数据, 或者返回 null 空数据库.

作为一个例子, name==’mongo’ 可以用很多方法来实现, 可以用 next() 来循环游标(需要校验是否为null), 或者当做数组返回第一个元素.

但是用 findOne() 方法则更简单和高效:

> var mongo = db.things.findOne({name:"mongo"});
> print(tojson(mongo));
{"name" : "mongo" , "_id" : ObjectId("497cf60751712cf7758fbdbb")}
>

findOne 方法更跟 find({name:”mongo”}).limit(1) 一样.

limit() 查询

你可以需要限制结果集的长度, 可以调用 limit 方法.

这是强烈推荐高性能的原因, 通过限制条数来减少网络传输, 例如:

> db.things.find().limit(3);
in cursor for : DBQuery: example.things ->
{"name" : "mongo" , "_id" : ObjectId("497cf60751712cf7758fbdbb")}
{"x" : 3 , "_id" : ObjectId("497cf61651712cf7758fbdbc")}
{"x" : 4 , "j" : 1 , "_id" : ObjectId("497cf87151712cf7758fbdbd")}
>