Published: Apr 18, 2021 by Jesús López-González

About this series

This series of posts describes our experience while learning q and kdb+ by Kx Systems. According to the wikipedia,

“kdb+ is a high-performance column-store database that was designed to process and store large amounts of data… created with financial institutions in mind… At the core of kdb+ is the built-in programming language, q, a concise, expressive query array language, and dialect of the language APL”

Although q is also a functional programming language, it has many features that make it different from other conventional languages such as Haskell, OCaml, etc. Given this situation, we’ll try to provide an overview of q basics, connecting the missing pieces to our previous knowledge on functional programming, using Scala and Spark to guide the explanations. We hope it also works in the opposite direction, so q programmers can also benefit from this introduction, which will be divided into the following posts:

1. Q as an (impure) functional language
2. Q as an array processing language
3. Q as a query language for kdb+

The first post introduces q as a functional language, showing the main q features that are already familiar to the conventional functional programmer which could be used from the very first day. The second post will put focus on q as an array processing language, which is probably the main source of q weirdness, so we’ll try to connect it to existing theory on the functional paradigm. Finally, the last post will introduce q and kdb+ as a query language and a column-oriented database, respectively, to make the data engineer happy.

Why q?

Each day, we set aside a time to experiment with new technologies, and we are especially interested on functional languages. A colleague from the Scala community pointed us towards q. The following items summarise the alleged language benefits why we decided to give it a go:

• Q is fast, sooo fast
• Q is a functional query language
• Q is a well-founded language that relies on APL
• Q is highly demanded in the financial industry
• Q is quite a challenge

After a few months of reading q material and coding, we can confirm that none of the previous items is a myth. Although we feel that we still have a long way to master q/kdb, we are confident that we have now a good perspective on how hard it is to learn this language. In fact, reflecting this experience is perhaps the major contribution of this series of posts. Having said so, we’re ready to go now!