%22%3E){kind=link}
What is a Graph Database?
NoSQL Graph Databases in a Nutshell
One does not always necessarily come into contact with graph databases. But I think as a data engineer or scientist, for example, it never hurts to know the basics in case you do encounter a new project involving graph databases.
Graph Databases vs. (Traditional) Relational Databases
The graph database uses mathematical graph theory and stores data and their relationships using nodes and edges. Graph algorithms can be used to execute complicated database queries at high speed. Graph databases have a different operating principle compared to relational databases [1]. It is possible to map graphs in relational databases. For the example of a social network, a table for the people and a table for relationships for the edges would be the solution.
Graph Databases Theory
In a graph database, both nodes and edges can store information. Special data structures like adjacency or index matrices are used for this purpose. In contrast to relational databases, complex relationships of different objects can be stored and processed in a graph database without additional auxiliary procedures. Edge weighting also makes it possible to assign a specific value to each relationship to express its strength or quality, for example, a distance between two locations.
Example with Neo4j and Cypher
Curious and want to try a round of the whole thing? With Neo4j a graph database solution, we can test the whole thing in a practical sandbox. Cypher is a graph query language that is used to query the Neo4j Database. Just like you use SQL to query a MySQL database, you would use Cypher to query the Neo4j Database[3].
As an example here, a movie database was used. It includes Movie, Person, Genre and Keyword nodes, connected by relationships as described in the following image:
Let’s start with a first query. I wanted to query all movies, where Tom Hanks is the actor.
MATCH (p:Person), (m:Movie)
WHERE p.name = "Tom Hanks"
RETURN p,mWith the clauses MATCH and WHERE I can realize that really easily and you maybe notice the similarity to SQL statements. Et voilà, the result is exactly what I expected:
Summary
Without a whole deep dive into the technology, I wanted to provide you at least with the most important technologies in a nutshell on this subject matter. You can also try out a first query yourself in the sandbox of Neo4j. As I said, not everyone will necessarily come into contact with graph databases, but I think it’s important to educate yourself continuously and especially as a person from the field of data it can’t hurt to get to know new concepts.
Summary and Further Readings
[1] Stefan Luber, Nico Litzel, Was ist eine Graphdatenbank? (2019)
[2] Microsoft, Create a graph database and run some pattern matching queries using T-SQL (2021)
[3] Neo4j, Getting Started with Cypher (2021)
[4] Cristina Escalante, https://neo4j.com/blog/node-js-react-js-developers-neo4j-movies-template/ (2016)