We have been investigating comparisons of performance and compression between the kdb+ binary format, Parquet and Arrow. Parquet is a columnar file format widely used in data warehousing, while Arrow is an on-disk format which is supported by some of the most popular data science tools. Some benchmarking has already been performed by the developers of Apache Arrow. We opted to use Python for …
AquaQ Coding Challenge: Columnar Transposition Ciphers
Encryption at its best is nearly unbreakable. At its most fun, it’s flimsy and thousands of years old. Our latest challenge is to break an existing columnar transposition cipher, and the challenge details are located here as challenge 35, “Columns“, on our challenge hub. Once you’ve gotten the spec, send us a function or script to find any codewords which …
AquaQuarantine Tech Talk Series 2021
We’ve decided to bring back our data-oriented talks for another run, with some help from our friends. Please use the sign up form below if you would like to join us. The content for our previous webinars is available here. Vaex, Arrow, Parquet Thursday 11th March, 4pm GMT Vaex is an open source python library for analysing large on-disk datasets …
Level 2 Storage Formats
Storing an order book can be tricky due to the large number of orders and prices that continually change throughout the day. Here we investigate and compare three common methods of storing an order book, assessing their benefits and limitations. To do that, the TorQ-CME pack was used to process CME MDP 3.0 FIX data on Soybean Futures for 10 …
Chinese New Year 2021!
At AquaQ we weren’t going to let lockdown number (I’ve lost count) stop us from celebrating Chinese New Year. This year is the ‘Year of the Ox’, said to bring stability, calmness, and great opportunities… I think we could all use a bit of that positivity! We’re counting on you, Ox. As we couldn’t celebrate Chinese New Year in person, …
A C++ Utopia – Some comments on interfacing kdb+ and C++
Recently quite a few projects at AquaQ have involved the manipulation of large numeric vectors in kdb+, and passing this data to and from external applications or tools written in C or C++. A practical example would relate to the use of machine learning libraries written in C++ that require output to be written to a kdb+ tickerplant or database. …
AquaQ Parquet kdb+ converter
Following on from the columnar and data formats blog post, we have decided to investigate creating a custom API for parquet. Traditionally, we would do this through python, however, we are interested to see what the performance benefits are. Kdb-Parquet is a library that can convert kdb+ tables to and from the Apache Parquet table format. The library provides a …
Real-time airline traffic dashboards with TorQ-Air
TorQ-Air is an application built on the TorQ framework that allows us to capture data through the Lufthansa developer API and perform both real-time and historical analytics on flight data. The solution includes a front-end built on Kx-dashboards, showing a number of real-time and historical stats. Download and installation instructions can be found on the TorQ-Air github repo. Founded in …
TorQ in 2020
2020 was a strange year for most of us. We’ve used it as an excuse to put a lot of work into extending TorQ- here’s a brief recap of some highlights. The first large contribution in 2020 was to provide the option of integrating Datadog into TorQ. Datadog is an extremely useful application monitoring service capable of tracking anything from …
AquaQ Food Challenge!
Here at AquaQ we’ve been coming up with ways to keep everyone engaged over lockdown, and what better way to do so than through the power of food! The first week was Italian week, so we sent out some ‘Dough it Yourself’ kits provided by Love Pizza Belfast, and challenged our team to make their best pizza. Our second week …