Getting Started With JanusGraph

Originally published at on February 25, 2021.

JanusGraph is a graph processing tool that can process graphs stored on clusters with multiple nodes. JanusGraph is designed for massive clusters and for real-time traversals and analytics queries.

In this post, we’ll look at a few queries that you would want to run the very first time you install JanusGraph and start playing with the Gremlin console. I did this just yesterday, so it’s still fresh in my memory. I saw the need to write this as I didn’t find a few of my questions answered in the official documentation, and I had to hunt down the information on the world wide web, which took some quality time. So, let’s get started.

Installing JanusGraph

To begin with, the official documentation is good enough for instructions on how to install and get the servers up. The issue I was talking about is that you need both Cassandra and Elasticsearch running to get the Gremlin server up. You can of course use alternatives to both Cassandra (such as HBase, Google Cloud Bigtable, Berkeley DB, or in-memory) and Elasticserach (such as Lucene, Solr, etc.), but I went the easiest route to get started with it.

The documentation says that you can skip mixed index backends (Elasticsearch, Solr, Lucene, etc.) and use only storage backends, but it doesn’t work. You need a mixed index backend along with a storage backend. So I had to download Elasticsearch as well and run it locally to get the Gremlin server up. This was a bit inconvenient, but no big deal, I guess. Now that we have all the moving parts moving as expected, let’s look at the data we’re going to play with.

The Sample Dataset

You can see that there are few special symbols in the graph. They have special meanings, of course.

Now that we have the graph and know what the different edges mean, let’s see if we can traverse the graph. To do that, we first need to get a starting point, which is one vertex. We can use the .has("property", "value") method on the graph to find a vertex with a particular value. I'll again go with the official example here.

We’ll start by getting the reference of the vertex with the “name” being “saturn.” For this, run the following query:

saturn = g.V().has('name', 'saturn').next()

Pretty straightforward. Now that we have the reference to the vertex in the saturn variable, let's see how we can traverse to Saturn's grandkid, Hercules. To find the grandson of Saturn, we first need to find the son of Saturn. From the graph, we can see that there's an incoming edge called "father" from Jupiter to Saturn. We can use this edge to traverse further. Because we have an incoming edge, we'll use the in() method on the saturn variable. Following is the query for this.

hercules = g.V(saturn).in('father').in('father')

The in() method accepts one parameter, which is the edge that has to be traversed. In this example, it's the 'father' edge. So from Saturn, we go to the next vertex which has a 'father' edge coming into Saturn, which in this case is Jupiter. From there, we again have the same traversal - in('father'). This will take us to the son of Jupiter, which is Hercules.

If you want to go the other way, that’s finding the grandfather of Hercules, you can run the following queries:

hercules = g.V().has('name', 'hercules').next() 
saturn = g.V(hercules).out('father').out('father')

As you can see, we’ve used the out() method here instead of in(). This is because we are traversing on the edge 'father' which is going out of the hercules vertex. It took me a while to understand when to use which method, but it's very logical once I figured it out.


hercules = g.V(saturn).repeat('father')).times(2).next() 
saturn = g.V(hercules).repeat(__.out('father')).times(2).next()

I don’t think it could be easier than this.

Selecting Multiple Vertices

g.V(hercules).out('father', 'mother').values('name')

So elegant.

As I mentioned at the beginning of this post, this is just getting started. There’s a lot more to JanusGraph and Gremlin. I’ll be posting more about these two as and when I learn new things. So watch out for this space if you are interested.

And if you like what you see here, or on my personal blog, and would like to see more of such helpful technical posts in the future, consider supporting me on Patreon and Github.

Become a Patron!

Coding, machine learning, reading, sleeping, listening, potato. and

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store