Operations on a Graph Database (Part 1 – Nodes)

Graph Database Tutorial

Part 1: Nodes

Part 2: Edges

Part 3: Types

Part 4: Properties

Part 5: Identifiers

Part 6: Traversals

Part 7: Sets

Part 8: Events

Graph databases are still quite unfamiliar to many developers. This is the first post in a series discussing the operations a graph database makes available to the developer. Just like there are only so many different things you can do on a relational database (like CREATE TABLE or INSERT), there are only so many things you can do on a graph database. It is worth looking at them one at a time, and that’s the goal of this series. This first post is on creating and deleting nodes.

To recap, a graph database contains nodes and edges, or MeshObjects and Relationships (as we call them in InfoGrid), or Instances and Links (as the UML would call them), or Resources and Triples (as the semantic web folks would call them), or boxes and arrows (as we draw them on a white board).

Nodes are those objects in a graph database that can stand on their own, they don’t depend on anything else. Edges are those objects that depend on the existence of (typically two) other objects, their source and their destination; we think of edges as connecting nodes.

To create a node in a graph database is one of its basic operations. For example, in InfoGrid, you can simply say:

MeshObject createMeshObject()

and voila, you have one. Similarly, you can delete a node by saying:

deleteMeshObject( MeshObject toDelete )

There are few conditions around those operations, such as that you have to have a transaction open, and that you have to have access rights to actually perform this operation, but that goes without saying.

When deleting a node, the graph database may require you to first delete all edges connected to the node before you get to delete it. Or, it may “ripple delete” all connected edges as part of the delete operation. There are some differences in the various graph database products on this; neither will make much of a difference to the developer.

If the graph database enforces a model (aka schema), as some graph databases do, you may need to make sure you don’t attempt to delete a node in a way that the schema would be violated. For example, if the schema says “an Order must be placed by exactly one Customer”, and you are attempting to delete the node representing the Customer, the graph database may prevent you from doing that as long as there still are nodes representing Order related to the Customer node. We’ll discuss schemas and graph databases in more detail in a later post.

For now, we learned two basic operations on a graph database:

  • create node
  • delete node.

Stay tuned for the next installment.


  1. Castle Hill Reptile Expo says on January 4th, 2011 at 9:15 pm:

    There are some graph databases on the market that require types for some elements (edges, for example) but don’t allow types for others (nodes, or properties), so hybrids are possible.

Comments are closed.