Put it differently, there are many ways to start exploring a database. You can learn about the paradigms behind that I think is the most important. It is essential to understand it on a high level how your data is stored and organized. For Neo4j I already covered this in the introduction to graph databases.
But I don’t think it’s important at the beginning of your journey to understand how the data is organized on a low level. What data structures are used in order to store tables, or documents, nodes, and relationships. How these structures point to each other. That can come later.
I think it’s better to jump right in where the fun is and start exploring your data and query it many different ways.
Then as a next step you can experiment with the different drivers from your preferred language(s) and after or in parallel with that you can explore the underlying systems from a closer aspect.
For Neo4j - the graph database of my choice - the query language is called Cypher. As I assume that if you are already interested in querying a graph database you know things about SQL. I will not be afraid of making connections between Cypher and SQL in this series.
I am going to write a series of articles about the concepts and different keywords of Cypher.
Let’s make a little recap before we start.
Neo4j is graph database implementing the labeled property graph data model. A lot of words, let’s cover them one by one:
Graph: It consists of nodes and relationships between the nodes. The relationships are always directed, they have a type, a start, and an end node. There are no broken relationships, that means each relationship must have both a start and an end point.
Property graph: A previously described graph can hold some attributes. An attribute means basically a key-value pair. Both nodes and relationship can have such attributes/properties. As examples, imagine Joe who has a birthday attribute:
{birthDay: 25/01/1985}
. And for a relationship, let’s say that there is a:MANAGES
relationship between Jim and Joe. The key of an attribute on:MANAGES
can besince
and the value is2017
.Labeled property graph: On the top of this, nodes can have labels. Labels are tags which can represent different roles within your domain. Jim can a have a label of
:EMPLOYEE
and R&D can be:DEPRATMENT
. A node can be tagged with multiple labels.
To finish this post, let’s see how easy it is to write down a relationship in Cypher.
Jim who is an employee is member of R&D department since 2017. The department has a size of 250 people.
(jim:EMPLOYEE {name: "Jim"})-[:MEMBER_OF {since: 2017}]->(rnd:DEPARTMENT {name: "R&D", size:250})
Some characteristics that you can see immediately:
- nodes are described between parentheses
- relationships are written between brackets
- labels of a node and the type of a relationship will come after a colon
- attributes are put between braces in a JSON style
- the direction of the relationship is visible through the direction of the arrow
Next time I’m going to introduce the CREATE
keyword to you so that we can create our first small graph.
Stay tuned and if you want to get notified of the latest articles follow me on Twitter!