sharing lagi… :p kali ini tentang pentingnya normalisasi database.
tapi maaf, kali ini tetep dibiarkan seperti sumbernya aja. daripada di translate dengan bahasa yang kacau…
Normalization calls for application of specific rules and design of your databases, wherein only relevant data is stored in tables. When the term database normalization is used, it means that data in a database has been processed and organized in an efficient manner. Normalization creates a relationship base between different tables, aimed at doing away with inconsistent dependencies and redundant occurrences of data.
Normalization creates a flexible database that will facilitate quick creation, searching and sorting of indexes. Tables will have more rows per page, but fewer columns and information in each row are directly related to a primary key. The indexes are more clustered and, therefore, are less rigid in facilitating query tuning and quick modification of data due to fewer numbers of indexes in each table. If your tables have only relevant data, then no redundancy exists, and execution of database triggers is fast.
Redundancy in databases occurs when you store the same data repeatedly in your database. For example, in an address table, you may note zip codes repeating for several addresses. This repetition of zip codes indicates that you should create a table specifically to store zip codes, and the data directly related to zip codes. The primary aim of normalization is to facilitate quick searching and indexing of data in the various tables within your database. Duplicated and inconsistent data slows the querying and indexing processes and may cause inaccurate reporting of your information.
Inconsistent dependency occurs when you store data related to a piece of data in a table, but not necessarily related to the primary key for that table. For optimum performance of any database, you must normalize the database by organizing the data in a systematic manner, storing only data that is directly related to the primary key in a given table. For example, a zip code table would have each zip code listed once in each row, and the matching area in the same row, though in a different column.
Databases are normalized based on certain design rules; thus, a database is said to be normalized if the first normal form is satisfied. The first normal form establishes the fundamental rules for database organization, and at this level, you remove repetitive data and place that data in individual tables with each row classified by a primary key. The second normal form eliminates subsets of data, or redundant data applicable to multiple rows in the same table, and houses them in different tables. With the help of a foreign key, the second normal form links the new tables with the previous ones. The third normal form removes redundant, dependent columns not related to the primary key. Although fourth and fifth normal forms exist, they are not often used in database designs.
sumber : http://www.ehow.com/info_8120428_normalization-must-database-application-business.html
untuk info terkait lainnya bisa kunjungi :