This post explains what database normalization consists of ,
why it is necessary and what types of normalization exist.
There are some rules that must be taken into account when
making a database. Based on the above, the following formulation can be given: ▷
database normalization ◁ is a process of organizing data
in a certain way and design guidelines. In other words, the tables and the
relationships between them (relationships) are created according to the rules.
In this way, the necessary level of data security is provided and the database
itself becomes more flexible. Also, inconsistent dependencies and redundancy
are eliminated.
Advantages of standardization:
Standardization is not mandatory, but it does provide the
following benefits:
• The sampling process is simplified . It is about
simplifying the query work, that is, the user will be able to retrieve the
desired information with relatively simple queries.
• The integrity of the data is guaranteed . One can talk
about minimizing the distortion of information and reducing the probability of
data loss.
•Scalability is improved . If the normalization rules are
observed, favorable conditions are formed for the growth of the database.
• There is no data redundancy . Redundancy is a notorious
problem of unproductive use of free hard disk space that makes database
maintenance difficult. In some cases, this problem is compounded by the fact
that if it is necessary to alter records of the same type of data stored in multiple
places (tables), the user will have to perform the alterations everywhere,
which is quite a laborious task. . It is much easier to make, for example, data
about cities stored only in the Cities table and nowhere else. In short,
redundancy implies duplication, which not only complicates the database, but
also increases its size.
• Absence of inconsistent dependencies . Inconsistent
dependencies prevent access to data, because the path to that information can
be wrong and illogical. In the Cities table it is logical to find the cities,
the number of inhabitants, etc., but not the addresses and names of the
inhabitants (for this information we need another table): Citizens.
Normalization is carried out:
To bring the database to a normal state, you need:
1. Merge existing data into groups .
2. Clarify the logical relationships between the groups . To
ensure the accuracy of the links, the fields that are linked must be of the
same type.
If the table is not normalized, you can store the
information for multiple entities and include the repeating columns which in
turn can store the duplicate values. However, if normalized, each table stores
information about a single entity.
The normalization involves the use of normal forms with
respect to the structure of the data available. There are several normalization
rules. Each of these is called a normal form (FN). Each of these shapes, except
the first, assumes that the previous normal shape has already been applied to
the data. When the first rule is executed, the database is represented in the
first normal form (1FN), and when all three rules are executed it is
represented in the third normal form (3FN).
There are seven of these forms (levels), but in practice for
most applications it is sufficient to normalize the database to the third
normal form (strictly speaking, the database will be considered normalized when
a 3FN or higher).
True, it is not always feasible to offer full compliance
with standards and specifications because standardization will require the
creation of additional tables and it is not always acceptable or not accepted
by customers. But if you have to break the rules, you have to understand that
all related issues, including uncoordinated dependencies and redundancy, will
be taken into account and that this is acceptable to the application without
breaking its functionality.
