Introduced in SQL Server 2017, the graph database stores nodes and edges within SQL Server, allowing you to use T-SQL to query graph data. Graph data is specifically oriented around many-to-many relationships. Graph data doesn’t lend itself to traditional relational structures but instead is focused on hierarchies and relationships past those more traditional structures. Graph data is separated from either traditional analytical data, better served by columnstore indexes, or OLTP systems, better served by rowstore indexes. However, when you have a data set that consists of complex hierarchical data, focused on the relationships between that data, the graph database comes into play.