Estimated reading time: 16 minutes

Together with my team, I took a deep dive into the available fuzzy search approaches and algorithms for quite a while, in order to find a performant solution for the various projects ArangoSearch gets used for. 

Since the introduction of ArangoSearch back in 2018, many of our users have asked for fuzzy search support. We worked hard on getting this done, and the whole team is now excited to finally make fuzzy search available with the upcoming ArangoDB 3.7.

In the following, we will share our learnings and hope they are useful for you.

What is Fuzzy..

In this article, we’re going to dive deeper into our recently released feature preview in Milestone ArangoDB 3.4 – ArangoSearch which provides a rich set of information retrieval capabilities. In particular, we’ll give you an idea of how our search engine works under the hood.

For the upcoming ArangoDB 3.4 release we’ve implemented a set of information retrieval features exposed via new database object `View`. The `View` object is meant to be treated as another data source accessible via AQL and the concept itself is pretty similar to a classical “materialized” view in SQL.

While we are still working on completing the feature, you can already try our retrieval engine in the Milestone of the upcoming ArangoDB 3.4 released today. You can download and give it a spin here:

Milestone ArangoDB 3.4