Database system principle Fragment translation

RT, in order to improve their knowledge of the database of CI Knowledge hui Reserve, I decided to do a little translation, as a practice.

Introduction to the first part

A database management system (DBMS) is a collection of interrelated data and a set of programs for accessing those data. A data collection, usually a database, that contains data related to the enterprise. The basic goal of the DBMS is to provide a convenient and efficient way to store and retrieve database information.

The database management system is designed to manage a large amount of information. The management of data involves defining the storage structure of information and the mechanism of providing operational data. In addition, the database system must ensure the security of the data store, even if the system crashes or attempts unauthorized access. If the data is shared among several users, the system must avoid possible exception results.

Because information is important in most organizations, computer scientists have developed a wealth of concepts and techniques for managing data. These concepts and techniques make up the focus of the book. This chapter provides a brief introduction to the principles of the database system.

　　

Part II relational database

The data model is a conceptual tool for describing data, data relationships, data semantics, and consistency constraints. In this section, we focus on the relational model. We will study the relationship model in chapter two, which uses a table set to represent the relationship between data and data. The concept of simplicity makes the relational database widely adopted; Today most database products are based on relational models. The relational model describes the database at the logical and view level, abstracting the data storage details at the lower levels. The entity-relationship model, which will be discussed in chapter seventh (Part II), is a higher-level relational model that is widely used in database design.

In order for the data in the relational database to be available to the user, we have to solve some problems. The most important issue is how users specify requests to retrieve and update data, and several query languages have been invented to perform the task. The second and most important point is the integrity and protection of data. The database needs to protect data from user actions, whether intentional or unintentional.

The third, fourth, and fifth chapters cover the contents of the SQL language, which is the most commonly used query language today. Chapters III and fourth provide introductory and intermediate-level SQL language descriptions. The fourth chapter also contains the mandatory integrity constraint and permission check mechanism of the database, which control the access and update behavior that the user will implement. The fifth chapter contains more advanced topics, including access to SQL in programming languages and data analysis using SQL.

The sixth chapter covers three forms of query language, relational algebra, tuple relationship calculus and domain relational calculus, which are three kinds of declarative query language based on mathematical logic. These formal languages are based on SQL. There are two other user-friendly languages, QBE and Datalog, which are described in Appendix B (available through db-book.com online access).

Part III data storage and querying

Although the database system provides a high-level view of the data, the data must eventually be stored as bits on one or more devices. The vast majority of today's databases store data on disk (and more and more SSDs) and crawl data into memory for processing, or replicate data on tapes and other backup devices. The physical characteristics of a storage device play a major role in how data is stored, especially because random access to a piece of data on a hard disk is much slower than doing the same thing in memory: Disk access takes dozens of milliseconds, and memory access requires only a few decades of subtlety.

The tenth chapter, from the physical storage that is not answered, contains a mechanism for reducing data loss due to device failure. The next chapter describes how records are mapped to files, which are mapped sequentially to the hard disk.

Many queries refer only to a small subset of records in a file. An index is a structure that helps you quickly locate the relationships of records you need, and it doesn't require you to check all the records. The index examples in this textbook are not indexed in the database, but are used for human use. The 11th chapter describes the types of indexes used in the centralized database.

The user query must run on the database content and the contents of the database reside on the storage device. Typically, the query is split into smaller operations, the corresponding relational algebra operations. The 12th chapter describes the way queries are handled, the implementation of individual manipulation algorithms, and an overview of how these operations are performed synchronously to handle queries.

When processing queries, there are many ways to choose from, and their costs are quite different. The usefulness of query optimization is to find the minimum cost when evaluating a given query. The 13th chapter describes the working process of query optimization.

Fourth part

The term "transaction" refers to a single logical unit formed by a collection of operations in the work. For example, the act of transferring money from one account to another is a transaction with two updates, one for each account.

It is important that all behaviors of a transaction are fully executed, or that the effect of each incomplete transaction part can be revoked in order to prevent some failures. This attribute is called "atomicity". Further, once a transaction executes successfully, his effect must be maintained in the database-a system error should not cause the database to forget the transaction that was successfully executed completely. This attribute is made persistent.

In a database system that is currently executing multiple transactions, if the behavior of updating shared data is not controlled, there is a potential possibility for the transaction to see inconsistent intermediate states caused by updates created by other transactions. Such a scenario could result in an incorrect update of the data stored in the database. Therefore, the database system must provide a mechanism to make each transaction unaffected by the effects of other executing transactions. This attribute is called isolation.

The 14th chapter describes the details of the transaction principle, including atomicity, persistence, isolation, and other attributes provided by the transaction abstraction. In particular, this chapter describes the principle of isolation precisely by the concept of being called serializable.

The 15th chapter describes several concurrency control techniques used to achieve isolation. The 16th chapter describes the recovery management component of the database, which implements the properties of atomicity and persistence.

In general, the transaction management component of the database system allows application developers to focus on their own transactional implementations, ignoring concurrency and fault tolerance issues.

Database system principle Fragment translation