Data redundancy can cause data anomalies in a database - most commonly insertion, deletion and update errors. The process of data normalization helps to eliminate data redundancy and its resultant anomalies.
Data redundancy occurs when a specific piece of data can be found in more than one area of the database. A common example would be a university or college's database of current enrollment in courses. One student might be enrolled in several different courses, so their individual student record may be reproduced several times. In a manufacturing scenario, a single vendor may be used for various projects and products.
It is important to eliminate the occurrence of data redundancy while still maintaining data integrity through the multi-step normalization process. Redundancy typically results in three common data anomalies - or instances where the data is inconsistent.
- Insertion anomaly - where data cannot be stored or updated unless another piece of data is stored at the same time. In the student records example, this might occur if it is not possible to enter a student record until they have enrolled in a course.
- Update anomaly - where one of the copies of a record is updated while the other is not; all copies should be updated simultaneously.
- Deletion anomaly - this occurs when deleting one piece of data means that other information is lost as well. In the student examples, this might occur if deleting a course meant also deleting the related student records.