What is database normalization in simple words?

Question

Database

After reading many articles, and even a couple of chapters of books, I still did not understand what normalization is. In different books / articles / videos, they usually say that only 3 levels of normalization are enough, while the criteria for these levels are different for different sources, as I understand it, data normalization is needed to eliminate redundancy (repetition of data) and, if we go from the opposite, then data that is repeated / may be repeated in the future, that is, not unique in the table, should be placed in a separate entity / table and relationships should be established between them, but in most examples there are duplicate values ​​in the database, the authors say that the values ​​\u200b\u200bshould not be repeated and at the same time leave duplicate values. ... I'm completely confused, please help, and sorry in advance for the stupid question)

Answer 1

The database is needed to store statements about the state of the information system.
We consider a situation in which there are more than one statements about the state of some part of the system as duplication of information. (You have this situation is called "redundancy").
If there is duplication of information, then a situation is possible when the statements may contradict each other. In this case, the database is inconsistent and the information system may not work correctly.
Normalization of the database - elimination of duplication of information in the database in order to achieve consistency, in the first place. Normal forms are the steps you need to take on the path to consistency (do it once, do it twice...)
A side effect of normalization is to reduce the size of the database.
In cases where performance is more important than consistency, the database is denormalized (introducing redundancy) for certain queries in order to avoid table joins (JOINs).