1、Unit 6 Database第六單元： 數(shù)據(jù)庫(kù) Section A:課文A：Database Overview數(shù)據(jù)庫(kù)概覽I. Introduction一、引言 Data storage traditionally used individual, unrelated files, sometimes called flat files.數(shù)據(jù)存儲(chǔ)傳統(tǒng)上是使用單獨(dú)的沒有聯(lián)系的文件，這些文件有時(shí)稱為平面文件。In the past, each application program in an organization used its own file.在過去，一個(gè)機(jī)構(gòu)中的每個(gè)應(yīng)用程序都使用自己

2、的文件。In a university，for example, each department might have its own set of files:例如，在一個(gè)大學(xué)中，每個(gè)部門都可能有其自己的文件集：the record office kept a file about the student information and their grades, the financial aid office kept its own file about students that needed financial aid to continue their education，the

3、 scheduling office kept the names of the professors and the courses they were teaching, the payroll department kept its own file about the whole staff (including professors), and so on.檔案辦公室保存著關(guān)于學(xué)生信息和學(xué)生成績(jī)的文件；經(jīng)濟(jì)資助辦公室保存著其自己的關(guān)于需要經(jīng)濟(jì)資助以繼續(xù)學(xué)業(yè)的學(xué)生的文件；調(diào)度辦公室保存著教授的姓名和他們所教的課程；工薪發(fā)放部門保存著其自己的關(guān)于全體教職員工（包括教授）的文件，等等。To

4、day, however, all of these flat files can be combined in a single entity, the database for the whole university.然而，所有這些平面文件今天都可結(jié)合成一個(gè)實(shí)體供整個(gè)大學(xué)使用的數(shù)據(jù)庫(kù)。Although it is difficult to give a universally agreed definition of a database, we use the following common definition：a database is a collection of relat

5、ed, logically, coherent data used by the application programs in an organization.雖然難以給出一個(gè)普遍接受的數(shù)據(jù)庫(kù)定義，但我們使用下面常見的定義： 一個(gè)數(shù)據(jù)庫(kù)是被一個(gè)機(jī)構(gòu)內(nèi)的應(yīng)用程序所使用的具有邏輯相干性的相關(guān)數(shù)據(jù)的集合。 Database Management Systems二、數(shù)據(jù)庫(kù)管理系統(tǒng) A database management system (DBMS) defines, creates, and maintains a database.數(shù)據(jù)庫(kù)管理系統(tǒng)定義、創(chuàng)建和維護(hù)數(shù)據(jù)庫(kù)。The DBMS also

6、 allows controlled access to data in the database.數(shù)據(jù)庫(kù)管理系統(tǒng)也允許對(duì)數(shù)據(jù)庫(kù)中的數(shù)據(jù)進(jìn)行受控的訪問。A DBMS is a combination of five components：hardware, software, data, users, and procedures.一個(gè)數(shù)據(jù)庫(kù)管理系統(tǒng)由5個(gè)組成部分構(gòu)成：硬件、軟件、數(shù)據(jù)、用戶和規(guī)程。1Hardware1、硬件 The hardware is the physical computer system that allows access to data.硬件是指允許訪問數(shù)據(jù)的計(jì)算

7、機(jī)物理系統(tǒng)。For example，the terminals, hard disk，main computer and workstations are considered part of the hardware in a DBMS.例如，終端、硬盤、主機(jī)和工作站被認(rèn)為是數(shù)據(jù)庫(kù)管理系統(tǒng)的硬件組成部分。2Software2、軟件 The software is the actual program that allows users to access，maintain, and update data.軟件是指允許用戶訪問、維護(hù)和更新數(shù)據(jù)的實(shí)際程序。In addition, the so

8、ftware controls which user can access which parts of the data in the database.另外，軟件還控制著哪個(gè)用戶可以對(duì)數(shù)據(jù)庫(kù)中的哪部分?jǐn)?shù)據(jù)進(jìn)行訪問。3Data3、數(shù)據(jù) The data in a database is stored physically on the storage device.數(shù)據(jù)庫(kù)中的數(shù)據(jù)存儲(chǔ)在物理存儲(chǔ)設(shè)備上。In a database, data is a separate entity from the software that accesses it.在一個(gè)數(shù)據(jù)庫(kù)中，數(shù)據(jù)是獨(dú)立于對(duì)其進(jìn)行訪問的

9、軟件的一個(gè)實(shí)體。This separation allows an organization to change the software without having to change the physical data or the way in which it is stored.這種獨(dú)立使一個(gè)機(jī)構(gòu)可以在不必更改物理數(shù)據(jù)及其存儲(chǔ)方式的情況下更改軟件。If an organization decides to use a DBMS，then all the information needed by the organization should be kept together as

10、 one entity, to be accessible by the software in the DBMS.如果一個(gè)機(jī)構(gòu)決定使用一個(gè)數(shù)據(jù)庫(kù)管理系統(tǒng)，那么該機(jī)構(gòu)所需要的所有信息都應(yīng)作為一個(gè)實(shí)體保存在一起，可由數(shù)據(jù)庫(kù)管理系統(tǒng)中的軟件訪問。4. Users4、用戶The term user in a DBMS has a broad meaning.在數(shù)據(jù)庫(kù)管理系統(tǒng)中，用戶這個(gè)術(shù)語有著廣泛的定義。We can divide users into two categories:我們可以將用戶分為兩類：end users and application programs.最終用戶和應(yīng)用程序。En

11、d users are those humans who can access the database directly to get information.最終用戶是指可直接訪問數(shù)據(jù)庫(kù)以獲取信息的人。There are two types of end users:最終用戶又分為兩類：database administrators (DBAs) and normal users.數(shù)據(jù)庫(kù)管理員和普通用戶。Database administrators have the maximum level of privileges and can control other users and t

12、heir access to the DBMS, grant some of their privileges to somebody else, but retain the ability to revoke them at any time.數(shù)據(jù)庫(kù)管理員擁有最高程度的特權(quán)，可以控制其他用戶及其對(duì)數(shù)據(jù)庫(kù)管理系統(tǒng)的訪問，可以將其某些特權(quán)授予其他人并保留隨時(shí)收回這些特權(quán)的能力。A normal user, on the other hand，can only use part of the database and has limited access.另一方面，普通用戶只能使用數(shù)據(jù)庫(kù)的一部分

13、，只能進(jìn)行有限的訪問。The other users of data in a database are application programs.數(shù)據(jù)庫(kù)中數(shù)據(jù)的其他用戶就是應(yīng)用程序。Applications need to access and process data.應(yīng)用程序需要訪問和處理數(shù)據(jù)。For example, a payroll application program needs to access part of the data in a database to create paychecks at the end of the month.例如，工薪發(fā)放應(yīng)用程序需要在月

14、底訪問數(shù)據(jù)庫(kù)中的部分?jǐn)?shù)據(jù)，來開支付工薪的支票。5Procedures5、規(guī)程 The last component of a DBMS is a set of procedures or rules that should be clearly defined and followed by the users of the database.數(shù)據(jù)庫(kù)管理系統(tǒng)的最后一個(gè)組成部分就是應(yīng)該明確定義并為數(shù)據(jù)庫(kù)用戶所遵循的一套規(guī)程或規(guī)則。 . Database Architecture三、數(shù)據(jù)庫(kù)體系結(jié)構(gòu)The American National Standards Institute Standards

15、 Planning and Requirements Committee (ANSI/SPARC) has established a three-level architecture for a DBMS:美國(guó)國(guó)家標(biāo)準(zhǔn)協(xié)會(huì)標(biāo)準(zhǔn)計(jì)劃與需求委員會(huì)（ANSI/SPARC）為數(shù)據(jù)庫(kù)管理系統(tǒng)確立了一個(gè)包含3個(gè)層次的體系結(jié)構(gòu)：internal, conceptual, and external (Figure 6A-1) .內(nèi)層、概念層和外層（圖6A-1）。Figure 6A-1:圖6A1：Database Architecture數(shù)據(jù)庫(kù)體系結(jié)構(gòu)1. Internal Level1、內(nèi)層The inte

16、rnal level determines where data is actually stored on the storage devices.內(nèi)層決定數(shù)據(jù)在存儲(chǔ)設(shè)備上的實(shí)際存儲(chǔ)位置。This level deals with low level access methods and how bytes are transferred to and from storage devices.該層涉及低級(jí)訪問方法，以及字節(jié)如何傳向和傳自存儲(chǔ)設(shè)備。In other words, the internal level interacts directly with the hardware.換

17、句話說，內(nèi)層直接與硬件交互。2. Conceptual Level2、概念層The conceptual level defines the logical view of the data.概念層定義數(shù)據(jù)的邏輯視圖。The data model is defined on this level, and the main functions of the DBMS, such as queries，are also on this level.數(shù)據(jù)模型在該層定義，數(shù)據(jù)庫(kù)管理系統(tǒng)的主要功能如查詢也在該層。The DBMS changes the internal view of data to

18、 the external view that users need to see.數(shù)據(jù)庫(kù)管理系統(tǒng)把數(shù)據(jù)的內(nèi)部視圖轉(zhuǎn)化為用戶需要看到的外部視圖。The conceptual level is an intermediary and frees users from dealing with the internal level.概念層是中介層，它使得用戶不必與內(nèi)層打交道。3. External Level3、外層The external level interacts directly with the user (end users or application programs).外層直接

19、與用戶（最終用戶或應(yīng)用程序）交互。It changes the data coming from the conceptual level to a format and view that is familiar to the users.它將來自概念層的數(shù)據(jù)轉(zhuǎn)化為用戶所熟悉的格式和視圖。IV. Database Models四、數(shù)據(jù)庫(kù)模型A database model defines the logical design of data.數(shù)據(jù)庫(kù)模型定義數(shù)據(jù)的邏輯設(shè)計(jì)。The model also describes the relationships between different

20、parts of the data.它也描述數(shù)據(jù)的不同部分之間的關(guān)系。In the history of database design，three models have been in use:在數(shù)據(jù)庫(kù)設(shè)計(jì)史上，使用過3種數(shù)據(jù)庫(kù)模型：the hierarchical model, the network model，and the relational model.層次模型、網(wǎng)絡(luò)模型和關(guān)系模型。1. Hierarchical Database Model1、層次數(shù)據(jù)庫(kù)模型In the hierarchical model, data is organized as an inverted

21、tree, each entity has only one parent but can have several children.在層次模型中，數(shù)據(jù)被組織成一棵倒置的樹。每個(gè)實(shí)體只有一個(gè)父，但可有數(shù)個(gè)子。At the top of the hierarchy, there is one entity, which is called the root.在分層結(jié)構(gòu)的頂部，有一個(gè)實(shí)體，稱為根。Figure 6A-2 shows a logical view of an example of the hierarchical model.圖6A-2給出了一個(gè)層次模型例子的邏輯視圖。The hi

22、erarchical model is now obsolete.層次模型現(xiàn)在已經(jīng)過時(shí)。Figure 6A-2:圖6A-2：An Example of the Hierarchical Model Representing a University代表大學(xué)層級(jí)的一個(gè)層次模型實(shí)例2. Network Database Model2、網(wǎng)絡(luò)數(shù)據(jù)庫(kù)模型In the network model, the entities are organized in a graph, in which some entities can be accessed through several paths (Figu

23、re 6A-3).在網(wǎng)絡(luò)模型中，實(shí)體以圖的形式來組織，圖中的有些實(shí)體可通過多條路徑訪問（圖6A-3）。There is no hierarchy.網(wǎng)絡(luò)模型沒有分層結(jié)構(gòu)。This model is now also obsolete.這種模型現(xiàn)在也已經(jīng)過時(shí)。Figure 6A-3:圖6A-3：An Example of the Network Model Representing a University代表大學(xué)的一個(gè)網(wǎng)絡(luò)模型實(shí)例3. Relational Database Model3、關(guān)系數(shù)據(jù)庫(kù)模型In the relational model, data is organized in t

24、wo-dimensional tables called relations.在關(guān)系模型中，數(shù)據(jù)被組織成稱為關(guān)系的二維表。There is no hierarchical or network structure imposed on the data.關(guān)系模型沒有分層或網(wǎng)絡(luò)結(jié)構(gòu)強(qiáng)加于數(shù)據(jù)。The tables or relations are, however, related to each other (Figure 6A-4).然而，表或關(guān)系是相互關(guān)聯(lián)的（圖6A-4）。Figure 6A-4:圖6A-4：An Example of the Relational Model Repres

25、enting a University代表大學(xué)的一個(gè)關(guān)系模型實(shí)例The relational database management system (RDBMS) organizes the data so that its external view is a set of relations or tables.關(guān)系數(shù)據(jù)庫(kù)管理系統(tǒng)組織數(shù)據(jù)，使其外部視圖呈現(xiàn)為關(guān)系或表的集合。This does not mean that data is stored as tables:這并不意味著數(shù)據(jù)以表的形式存儲(chǔ)：the physical storage of the data is independe

26、nt of the way in which the data is logically organized.數(shù)據(jù)的物理存儲(chǔ)與數(shù)據(jù)的邏輯組織方式毫無關(guān)系。Figure 6A-5 shows an example of a relation.圖6A-5給出了一個(gè)關(guān)系的例子。A relation in an RDBMS has the following features:關(guān)系數(shù)據(jù)庫(kù)管理系統(tǒng)中的關(guān)系具有以下特征：Figure 6A-5:圖6A-5：An Example of a Relation關(guān)系實(shí)例Name. 名稱。Each relation in a relational database

27、should have a name that is unique among other relations.關(guān)系數(shù)據(jù)庫(kù)中的每個(gè)關(guān)系都應(yīng)具有一個(gè)名稱，而這個(gè)名稱在所有關(guān)系中是獨(dú)一無二的。Attributes. 屬性。Each column in a relation is called an attribute.關(guān)系中的每一列都稱為一個(gè)屬性。The attributes are the column headings in the table in Figure 6A-5.在圖6A-5的表中，屬性是列的標(biāo)題。Each attribute gives meaning to the data s

28、tored under it.每個(gè)屬性賦予存儲(chǔ)在其下面的數(shù)據(jù)以意義。Each column in the table must have a name that is unique in the scope of the relation.表中的每一列都必須具有一個(gè)在關(guān)系的范圍內(nèi)獨(dú)一無二的名稱。The total number of attributes for a relation is called the degree of the relation.一個(gè)關(guān)系的屬性總數(shù)稱為該關(guān)系的度。For example, in Figure 6A-5, the relation has a degr

29、ee of 3.例如，在圖6A-5中，關(guān)系的度為3。Note that the attribute names are not stored in the database:注意屬性名并不存儲(chǔ)在數(shù)據(jù)庫(kù)中：the conceptual level uses the attributes to give meaning to each column.概念層使用屬性給每一列賦予一定的意義。n Tuples. 元組。Each row in a relation is called a tuple.關(guān)系中的每一行稱為一個(gè)元組。A tuple defines a collection of attribu

30、te values.元組定義一組屬性值。The total number of rows in a relation is called the cardinality of the relation.一個(gè)關(guān)系中的總行數(shù)稱為該關(guān)系的基數(shù)。Note that the cardinality of a relation changes when tuples are added of deleted.注意一個(gè)關(guān)系的基數(shù)隨著元組的增加或刪除而改變。This makes the database dynamic.這使數(shù)據(jù)庫(kù)具有了動(dòng)態(tài)性。The relational model is one of th

31、e common models in use today.關(guān)系模型是今天使用的常見模型之一。The other two common models that are derived from the relational model are the distributed model and the object-oriented model.源自關(guān)系模型的另外兩種常見模型是分布式模型和面向?qū)ο竽Ｐ汀?. Distributed Database Model4、分布式數(shù)據(jù)庫(kù)模型The distributed database model is not a new model, but is b

32、ased on the relational model.分布式數(shù)據(jù)庫(kù)模型并非一種新模型，而是基于關(guān)系模型的。However, the data is stored on several computers that communicate through the Internet or a private wide area network.但是，數(shù)據(jù)存儲(chǔ)在通過因特網(wǎng)或?qū)Ｓ脧V域網(wǎng)通信的數(shù)臺(tái)計(jì)算機(jī)上。Each computer (or site) maintains either part of the database or the whole database.每臺(tái)計(jì)算機(jī)（或站點(diǎn)）保持?jǐn)?shù)據(jù)

33、庫(kù)的一部分或整個(gè)數(shù)據(jù)庫(kù)。In other words，data is either fragmented, with each fragment stored at one site，or data is replicated at each site.換句話說，數(shù)據(jù)或者是分段存儲(chǔ)的每個(gè)站點(diǎn)存儲(chǔ)一段，或者被每個(gè)站點(diǎn)復(fù)制一份。In a fragmented distributed database，data is localized - locally used data is stored at the corresponding site.在分段型分布式數(shù)據(jù)庫(kù)中，數(shù)據(jù)是本地化的，本地使用的數(shù)

34、據(jù)存儲(chǔ)在相應(yīng)的站點(diǎn)上。However, this does not mean that a site cannot access data stored at another site, but access is mostly local, but occasionally global.然而，這并不意味著一個(gè)站點(diǎn)不能訪問存儲(chǔ)在另一個(gè)站點(diǎn)上的數(shù)據(jù)，但訪問大多是本地性的，偶爾是全局性的。Although each site has complete control over its local data, there is global control through the internet

35、 or a wide area network.雖然每個(gè)站點(diǎn)對(duì)其本地?cái)?shù)據(jù)具有完全的控制，但也存在通過因特網(wǎng)或廣域網(wǎng)的全局控制。For example，a pharmaceutical company may have multiple sites in many countries.例如，一家制藥公司可能在許多國(guó)家擁有多個(gè)站點(diǎn)。Each site has a database with information about its own employees, but a central personnel department could have control of all the dat

36、abases.每個(gè)站點(diǎn)有一個(gè)數(shù)據(jù)庫(kù)，存儲(chǔ)著自己雇員的信息。但是，中心人事部門可以控制所有的數(shù)據(jù)庫(kù)。In a replicated distributed database，each site holds all exact replica of another site.在復(fù)制型分布式數(shù)據(jù)庫(kù)中，每個(gè)站點(diǎn)都有其他站點(diǎn)的一個(gè)完全副本。Any modification to data stored in one site is repeated exactly at every site.對(duì)一個(gè)站點(diǎn)所存儲(chǔ)的數(shù)據(jù)進(jìn)行的任何修改，都要在其他每個(gè)站點(diǎn)上精確地重復(fù)進(jìn)行。The reason for havi

37、ng such a database is security.擁有這種數(shù)據(jù)庫(kù)是為了安全。If the system at one site fails，users at the site can access data at another site.如果一個(gè)站點(diǎn)上的系統(tǒng)發(fā)生故障，該站點(diǎn)的用戶可以訪問另一個(gè)站點(diǎn)上的數(shù)據(jù)。5. Object-Oriented Database Model5、面向?qū)ο髷?shù)據(jù)庫(kù)模型The relational database has a specific view of data that is based on the nature of the database

38、s tuples and attributes.關(guān)系數(shù)據(jù)庫(kù)具有一個(gè)特定的數(shù)據(jù)視圖，該視圖基于數(shù)據(jù)庫(kù)元組與屬性的性質(zhì)。The smallest unit of data in a relational database is the intersection of a tuple and an attribute.關(guān)系數(shù)據(jù)庫(kù)中最小的數(shù)據(jù)單位是一個(gè)元組與一個(gè)屬性的交集。However, some applications need to look at data as other forms, for example, to see data as a structure, such as a r

39、ecord composed of fields.然而，有些應(yīng)用程序需要將數(shù)據(jù)視為其他形式，如看作一種結(jié)構(gòu)，像由字段構(gòu)成的記錄。An object-oriented database tries to keep the advantages of the relational model and at the same time allows applications to access structured data.面向?qū)ο髷?shù)據(jù)庫(kù)試圖保留關(guān)系模型的優(yōu)點(diǎn)，同時(shí)允許應(yīng)用程序訪問結(jié)構(gòu)化數(shù)據(jù)。In an Object-Oriented database, objects and their rel

40、ations are defined.在面向?qū)ο髷?shù)據(jù)庫(kù)中，對(duì)象及其之間的關(guān)系得到定義。In addition, each object can have attributes that can be expressed as fields.此外，每個(gè)對(duì)象可以具有可表示為字段的屬性。For example, in an organization, one could define object types for employee, department, and customer.例如，在一個(gè)機(jī)構(gòu)中，可以為雇員、部門和客戶定義對(duì)象類型。The employee class could defi

41、ne the attributes of an employee object (first name, last name, social security number, salary, and so on) and how they can be accessed.雇員類可以定義一個(gè)雇員對(duì)象的屬性（名、姓、社會(huì)保險(xiǎn)號(hào)碼、薪水等等），以及可以如何訪問它們。The department object could define the attributes of the department and how they can be accessed.部門對(duì)象可以定義部門的屬性，以及可以如何訪問它們。In addition, the database could create a relation between employee object and a department object to denote that the employee works in that department.此外，數(shù)據(jù)庫(kù)還可以在一個(gè)雇員對(duì)象與一個(gè)部門對(duì)象之間創(chuàng)建一種關(guān)系，以表示該雇員在該部門工作。. Database Design五、數(shù)據(jù)庫(kù)設(shè)計(jì)The design of any database is a lengthy and involved task that can onl