Вибір між реляційними та нереляційними базами даних

Creation of a large amount of data in industry, science and society has led to the problem of how to accommodate, store and process it. A separate issue is that this information is very variable, difficult to classify, and the speed of computers is becoming insufficient compared to its quantity.Therefore, effective organisation of data management and the choice of storage will affect the final result of their usage. The purpose of this article is to evaluate modern database models and their use for various sectors of society, to choose a more efficient and appropriate one depending on the advantages and disadvantages. Different types of databases are used to store data, the most common of which are relational SQL and non-relational NoSQL. SQL databases store data in the form of a set of rows and columns, i.e. in a tabular form.This has led to the emergence of non-relational NoSQL databases: key-value, document-oriented, columnar, and graph.The analysis showed that each type of database has its advantages and disadvantages. For example, SQL is a simple, mathematical logic-based, standardised and productive form. However, the disadvantages of databases are limitations in work-ing with unstructured data, the amount of which is growing, and scalability. NoSQLs do not store information in tabular form, have dynamic flexible schemas, and are capable of storing semi-structured and unstructured data. Compared to relational databases, they do not support complex queries, but this is compensated by their ability to store data of various types, support for horizontal scaling, and significant data processing speed. They are also capable of using aggregates and clusters. This makes it possible to use them widely for big data. Thus, we can conclude that the use of NoSQL databases is promising.