Introduction to Datalog

If you have mostly been working with “mainstream programming languages”, Eclair (and Datalog in general) might look foreign to you in terms of both the syntax and how these languages work under the hood. If this is the case, don’t worry! Datalog is a very minimal language and can be learned quickly.

Below are some links to introductory articles that explain some of the foundational concepts and ideas that Datalog is based on. One thing to note is that because there are several variants of Datalog, the syntax used in these articles might differ from Eclair’s syntax, but the ideas stay the same. If there is a big difference compared to how Eclair works, this will be pointed out.

• The Wikipedia entry for Datalog gives a good introduction to what Datalog is, the semantics of the language and how results are calculated. Note that Eclair handles queries differently: rules need to be marked as outputs for Eclair to compute and return the results.
• Introduction to Datalog is a great blogpost that explains many of the ideas in Datalog step by step with several examples. Keep in mind though that in Eclair facts and rules need to be explicitly marked if they are used as an input or output.
• Chapters 12 to 15 of Foundation of Databases (a.k.a. Alice’s book) focus on Datalog. It gives a more theoretical foundation and formalizes the concepts used in the language.
• Guide for writing Datalog fact extractors is a blogpost written by Luc Tielen, the creator of Eclair. The guide is focused on how to prepare a dataset for analysis with Soufflé Datalog, but the same approach works for Eclair as well.
• Static analysis using Haskell and Datalog is another tutorial on Luc’s blog. This post focuses again on doing analysis with Soufflé, but you can do the same for Eclair if you make use of the eclair-haskell library. (The code will almost be exactly the same on the Haskell side!)

Do you know of a good resource that is not in the list above? Submit an issue on Github and it will be added!