A database is large, integrated, structured collection of data, which is usually meant to model a real-world scenario or enterprise.
Definitions
- Record: A row in the database
- Attribute: A column in the database
Advantages Over A File System
File System (FS) | Database System (DS) |
---|---|
Program-data dependence: If the file structure changes, so does the program. What if you change data structure for one program? | Data independence: Separation of data and program, application logic. Central data repository, central management |
Duplication of data: Wasteful, inefficient, loss of data integrity | Minimal data redundancy: Data redundancy can be controlled (normalization) |
Limited data sharing: Data tied to application, hard/slow to create adhoc reports | Improved data sharing: Data is shared, a corporate resource, not a necessity for an application. External users can be allowed access Multiple views of data, arbitrary views of data |
Excessive program maintenance: Up to 80% of development time in traditional file based organisations is for maintenance | Reduced program maintenance: Data structure can change without application data changing |
Lengthy development times: Application has to do low level data management, figure out file format each time | **Novel ad hoc data access (no programming) :[[Structured Query Language |
DBMS (DataBase Management System)
See Database Management System
Relational Database
Transactional Database
#todo A transactional database is optimised for queries.
Informational Database
An informational database, is useful for aggregation and dimensional analysis.
- Good for numerical aggregation, such as with aggregate functions:
- How many?
- What is the average?
- What is the total cost?
- Useful for database dimensional analysis
Properties
Data Redundancy
Factors that increase data redundancy are:
- Increases when duplicate values exist
- Increased when derived attributes are stored as their own attribute.
Data redundancy can be reduced by normalisation
Data Integrity
- Increased by normalisation