精華區beta NTUMEB87-C 關於我們 聯絡資訊
Data Management Data management refers to the procedures used to acquire, access, and maintain data. Three major aspects of data management are data accuracy , data security, and data maintenance. Data accuracy (data integrity) means that the source of the data is reliable and the data is reported and entered correctly. Data security involves protecting data to keep it from being misused or lost. Data maintenance refers to the procedures used to keep data current. The Hierarchy of Data Data is organized in a hierarchy in which each higher level is made up of one or more elements from the lower level preceding it. A bit is a binary digit that usually is combined with other bits to represent data. A byte is a combination of eight bits that can represent a character. A field is a combination of one or more bytes and usually is the smallest item of meaningful data. A record is a group of related fields. A file is a collection of related records or data that is stored under a single name. A database is a group of related files. Sequential File Organization Files typically are organized based on the media on which they are stored or the manner in which they are processed. Sequential file organization, which is used with files stored on tape, means that records are stored sequentially, one after the other. Records are retrieved in the same order in which they are stored; the computer cannot retrieve a record without reading all the preceding records first. Indexed File Organization Indexed file organization uses a separate file, called an index, to record the location of a record on a storage device. An index consists of a list containing the values of one or more fields and the corresponding disk address for each record. Indexed files can be accessed sequentially and directly. With direct, or random, access, the system can go directly to a record without having to read the preceding records. Direct File Organization Direct file organization uses a record’s key value to determine the location on the disk where the record is stored. In a process called hashing, a formula is used to calculate the location of a record. Once a record is stored in its relative location within a direct file, it can be retrieved either directly or sequentially. Data Maintenance Data maintenance includes processes needed to keep data current and accurate. Records are added when additional data is needed to make the file current. Records are changed to correct inaccurate data and to update older data. Records are deleted when they are no longer needed. Database Management Systems A basic file-oriented system is made up of flat files, each of which is independent and contains all the information needed to process the records in that one file. A database management system (DBMS) is software that allows data and file relationships to be created, maintained, and reported. A database management system offers a number of advantages over a file-oriented system including reduced data redundancy (duplication), increased data accuracy, easier reporting, improved data security, and reduced development time. Databases are organized according to the following four types: hierarchical, network, relational, and object-oriented. Hierarchical and Network Databases In a hierarchical database, data is organized in a series like a family tree. Each parent record can have multiple child records, but each child record can have only one parent. All parent-child relationships are established when the database is created in a separate process that sometimes is called generating the database. A network database is similar to a hierarchical database except that each child record can have more than one parent. A child record is referred to as a member, and a parent record is referred to as an owner. Relationships can be established among different branches of the data, offering increased access capability for the user. Like a hierarchical database, however, data relationships must be established prior to the use of the database and must be redefined if fields are added or modified. Relational and Object-Oriented Databases In a relational database, data is organized in tables that are called relations. The tables are further divided into rows (called tuples) and fields (called attributes). The range of values that an attribute can have is called a domain. With a relational database, links between tables (files) can be established at any time, provided the tables have a common field. An object-oriented database keeps track of objects, which are entities that contain both data and the action that can be taken on the data. Object-oriented databases also are designed to store unstructured data such as photographs, video, and audio clips. PC Database Management Systems Popular PC database management systems include Access, Paradox, Approach, and FoxPro. Database management systems share a number of common features including a data dictionary, utilities, security, replication, recovery, and a query language. Query Languages A query language is a simple English-like language that can be used to specify the data seen on a report or screen display. Query-by-example (QBE) helps to construct a query by displaying a list of fields available in the files from which the query will be made. Three relational operations are used to query a relational database. The select relational operation picks certain records based on designated criteria. The project relational operation specifies fields that will appear on the query output. The join relational operation combines two files. One of the more widely used query languages is Structured Query Language (SQL). Dabase Administration The database administrator (DBA) is the person responsible for managing all database activities. The job of the DBA usually includes database design, user coordination, performance monitoring, system security, data distribution , and backup and recovery. Guidlines for Creating Database Files Carefully designed database files make it easier for a user to query a database and create reports. A process called normalization is used to organize the data in relational database files into the most efficient and logical file relationships. The following guidelines can be applied to both relational and flat files: design the file on paper first, include a unique key field, use separate fields for logically distinct items, do not create fields for information that can be derived from entries in other fields, allow enough space for each field, and set default values for frequently entered data. -- "Merciful and Helpful Goddess Office" Sorry for the late introduction,but my name is "Belldandy". I wish a goddess like you could stay with me forever ! -- ※ 發信站: 批踢踢實業坊(ptt.twbbs.org) ◆ From: chintei.m1.ntu.