In the recent years, the data landscape has changed significantly: massively structured and unstructured data is being generated from various sources. Applications and services are being developed with agile methodologies more frequently than ever, the data changes being fast and significant. Scalability requirements are also very important. Relational database management systems are not designed to handle all these challenges. Therefore, enterprises use NoSQL database solutions, which are more appropriate for new data requirements.

Graph databases belong to the NoSQL databases family. They have only existed for a few years and are used for storing and querying highly connected data, usually connected with many-to-many relations. SQL Server 2017 offers graph database capabilities in a feature called SQL Graph.

This chapter is divided into the following three sections:

• Introduction to graph theory and graph databases
• SQL Graph in SQL Server 2017
• Limitations in the current version

The first section starts with a brief introduction to graph theory. After a comparison between relational and graph tables, you will see examples and use cases for graph databases, and which solutions are available on the database market.

In the second section, you will see how SQL Server 2017 supports graph databases. You will learn how to create node and edge tables, how they are different from normal tables, and what the indexing recommendations are for them. Most of the time, you will traverse through graph structure; you will also learn how to do this in this section, with examples of usage of the new MATCH clause. At the end of the section, you can find an overview of built-in functions that deal with automatically generated columns in graph tables.

In the third section, you will see what is still missing in SQL Server 2017 and what would be nice to have in the set of graph capabilities in future versions.