3/6: Data Management, pt. 2 Refresh your memory Relational Data Model Hierarchical & Network Data Model Object-oriented DBMS Designing & Distributing a Database Trends Multidimensional Data Analysis, OLAP Hypermedia Linking databases to the web Data Warehousing & Data Marts Who deals with databases? Users & Administrators
Database Management System Software to create & maintain data Enables business applications to extract data Independent of specific computer programs
The 3 Key Components of a DBMS Data definition language: Defines data elements in database Data manipulation language: Manipulates data for applications SQL: Structured Query Language Emerging standard data manipulation language for relational databases Data dictionary: Formal definitions of all variables in database; Controls variety of database contents
Types of Databases: Relational Flexible, common format today. Data in table format Relation: table (sometimes called a file) Tuple: row (record) in table Field: column (attribute) in table
Relational DBs: Possible Relationships One-to-one One-to-many Many-to-many STUDENT ID CLASS STUDENT A B C CLASS 1 2 STUDENT A B C
Hierarchical Data Models Older, traditional system Top-level segment: Root. Parent-child relationships (one-to-many) ROOT FIRST CHILD 2nd CHILD Ratings Salary Compensation Job Assignments Pension Insurance Health Benefits Employer
Network Data Models A variation of the hierarchical data model Allows for many-to-many relationships STUDENT A B C CLASS 1 2
Goodbye, Hierarchical & Network Rarely, the network nor the hierarchical models are used for new databases now: they are legacy systems. Less flexible than relational DBMS Do not support ad-hoc queries Intensive programming required for changes to reports, queries, etc.
Object-Oriented DBMS Improvement over traditional DBMS to handle graphics, multimedia, etc. Variation: object-relational DBMS Hybrid of OODBMS & Relational DBMS
Designing a Database Conceptual (logical) design & physical design Abstract model of database from business perspective Deciding which data to include, exclude What to use as the key field? Normalization: creating small stable data structures from complex data Critical task: avoid redundancy. Choosing hardware & software for database
Designing a Database Conceptual (logical) design Entity-relationship diagram Illustrates relationship between various entities in a database 1 M ORDER CAN HAVE PART SUPPLIER ORDER: #, DATE, PART #, QUANTITY PART: #, DESCRIPTION, UNIT PRICE, SUPPLIER # SUPPLIER: #, NAME, ADDRESS
Distributing Databases: 2 Ways PARTITIONED: remote CPUs (connected to host) have files unique to that site, e.g., records on local customers DUPLICATE: each remote CPU has copies of common files, e.g., layouts for standard reports and forms Why distribute a database? Reduce vulnerability Increase service & responsiveness Can run on smaller, cheaper computers
Trends in DBMS Multidimensional Data Analysis Hypermedia Databases Comparing more than two aspects of data Synonym: On-line Analytical Processing (OLAP) Hypermedia Databases Storage and indexing of images, videos, sounds, multimedia, etc. Linking databases to the web http://mba.slu.edu/admissions.html
Data Warehouses & Data Marts Organization’s electronic library stores consolidated current & historic data for management reporting & analysis Provides a range of ad hoc & standardized query tools, analytical tools, and graphical outputs. Data Marts: smaller, specialized versions of the data warehouse customized for audience or task.
Data Warehouses & Data Marts
Who Deals With Databases? Users Administrators Define & organize database structure and content Develop security procedures Develop database documentation Maintain DBMS