MongoDB: Working with databases. Database device.

Preamble

The whole model of the database device in MongoDB can be presented as follows:

If in the relational databases the contents are tables, then in the working with  databases consists of collections.

Each collection has its own unique name – an arbitrary identifier consisting of no more than 128 different alphanumeric characters and an underscore.

In contrast to relational databases, MongoDB does not use a tabular device with a clearly defined number of columns and data types. MongoDB is a document-oriented system in which the central concept is a document.

A document may be represented as an object that stores some information. In a sense, it is similar to rows in relational subd, where rows store information about a single element. For example, it is a typical document:

{
"name": "Bill",
"surname": "Gates",
"age": "48",
"company": {
"name" : "microsoft",
"year" : "1974",
"price" : "300000"
}
}

The document represents a set of key-value pairs. For example, in the expression “name”: “Bill” name represents the key, and Bill – the value.

Keys represent rows. Values may differ by the type of data. In this case, almost all values also represent string type, and only one key (company) refers to a separate object. The following types of values are available in total:

• String: string data type, as in the above example (UTF-8 encoding is used for strings)
• Array: data type for storing element arrays
• Binary data: type for storing data in binary format
• Boolean: A boolean data type that stores logical TRUE or FALSE values, for example {“married”: FALSE}.
• Date: stores the date in Unix time format
• Double: numeric data type for storing floating-point numbers
• Integer: used to store integer values, for example {“age”: 29}.
• JavaScript: data type for storing javascript code
• Min key/Max key: used to compare values with the smallest/largest BSON elements
• Null: data type for storing the Null value
• Object: string data type, as in the above example
• ObjectID: data type for storing id document
• Regular: used to store regular expressions
• Symbol: data type, identical to a string. It is used mainly for those languages, which have special characters.
• Timestamp: used to store time

Unlike strings, documents may contain heterogeneous information.

For example, next to a document described above, one collection may contain another object, for example:

{
"name": "Tom",
"birthday": "1985.06.28",
"place" : "Berlin",
"languages" :[
"english",
"german",
"spanish"
]
}

It would seem different objects except for individual properties, but all of them may be in the same collection.

A couple of other important comments: in MongoDB queries are case sensitive and have strict typing. In other words, the next two documents will not be identical:

{"age" : "28" }
{"age" : 28}

If in the first case the age key is defined as a string value, in the second case the value is a number.

Document Identifier

For each document in MongoDB, there is a unique identifier called _id. When a document is added to the collection, this identifier is automatically created. However, the developer can explicitly specify the identifier himself, rather than relying on the automatically generated ones by specifying the corresponding key and its value in the document.

This field shall have a unique value within the collection. And if we try to add two documents with the same identifier to the collection, only one of them will be added, and when we add the second one we get an error.

If the identifier is not specified explicitly, MongoDB creates a special binary value with the size of 12 bytes. This value consists of several segments: the value of timestamp type with the size of 4 bytes, the machine identifier with 3 bytes, the process identifier with 2 bytes, and the counter with 3 bytes.

Thus, the first 9 bytes guarantee uniqueness among other machines, which may be replicas of the database. And the next 3 bytes guarantee uniqueness within one second for one process. Such a model of building an identifier ensures with a high probability that it will have a unique value because it allows you to create up to 16 777 216 unique ObjectId objects per second for one process.

Working with databases

About Enteros

IT organizations routinely spend days and weeks troubleshooting production database performance issues across multitudes of critical business systems. Fast and reliable resolution of database performance problems by Enteros enables businesses to generate and save millions of direct revenue, minimize waste of employees’ productivity, reduce the number of licenses, servers, and cloud resources and maximize the productivity of the application, database, and IT operations teams.