After a lot of planning, coding and testing we can finally share the first Release Candidate of ArangoDB 3.5 with you today. You can get it on our Technical Preview Download page: Tech Preview Community & Tech Preview Enterprise.
With this RC, we want to highlight two new features for all graph database enthusiasts: the new PRUNE Keyword & k Shortest Path Feature. As always, please note Release Candidates are for testing purposes only and should not be used in production. Please see limitations of RC1 at the bottom of the page!
If you want to share your feedback with us about the new features, please do so via GitHub (highlighting RC1).
Let’s dive in…
New PRUNE Keyword For Graph Traversal Syntax
The new PRUNE keyword in AQL can be very helpful in terms of query readability, but most importantly, for better performance when traversing a graph with active filter conditions. By setting PRUNE as your filtering method, the defined filter condition(s) get directly applied during traversal execution, reducing the overhead generated by a traversal query.
Please check out the video tutorial below showing the difference between FILTER and the new PRUNE with a hands-on example. You can also find more details in our docs.
New k Shortest Path Feature
In previous versions of ArangoDB, you could only search for a shortest path between a start and end vertex. With the new k Shortest Path feature you can now query for all shortest paths between two vertices, and sort your result set by path length or weight.
There are plenty of graph use cases where this can come in handy. A great example is navigation systems where your road or railway network is stored as a graph and you want to query for shortest travel time, distance or e.g. stations a train has to stop.
Take the new feature for a spin and find the k Shortest Path tutorial in our Training Center.
PRUNE and k Shortest Path are just two of the features we will announce with ArangoDB 3.5. Stay tuned for more updates and we highly appreciate any feedback about your experience during testing.
Limitations of RC1
A release candidate is of course still work in progress and some limitations exist. If you plan to upgrade any existing deployment of a GA version to a Release Candidate, we strongly recommend to always backup your systems first!
Known issues of RC1:
- For everybody who has a view of type arangosearch created and wants to upgrade ArangoDB from 3.4.x to RC1 of ArangoDB 3.5, we recommend to delete the view in 3.4.x, then upgrade to RC1 and create the view anew.
Other known issues can be found in the devel docs.