Technology

This week we’ve been looking at joining two huge tables in Spark into a single table. It turns out that it is not a straightforward exercise to join data based on an array of IDs. In this blog I’ll show one way of doing this.

My Synology disk crashed and so did my Docker set up. Basically, the CI/CD pipeline for my programs no longer existed. The wonderful thing of an awful crash like this, is that I could rethink my setup. The result is what I would call “a poor man’s CI/CD”. It’s just Git, Docker, Docker Compose and Cron. It is easy to set up and it might be all you need.

For some blogs I need to capture terminal screens. The recording of these types of screens have different requirement then normal application or website recordings. The bottom of the video is the most important part, the letters need to be crisp and readable for the end user.

Our data strategy specifies that we should store data on S3 for further processing. Raw S3 data is not the best way of dealing with data on Spark, though. In this blog I’ll show how you can use Spark Structured Streaming to write JSON records of a Kafka topic into a Delta table.

There is a lot of code that needs to make a selection based on a maximum value. One example are Kafka reads: we only want the latest offset for each key, because that’s the latest record. What is the fastest way of doing this?

Tired of the dull Python syntax highlighting in Databricks? Just copy this code into your Magic CSS editor, change it (to your own style), pin it & enjoy!

At Wehkamp we use Apache Kafka in our event driven service architecture. It handles high loads of messages really well. We use Apache Spark to run analysis. From time to time, I need to read a Kafka topic into my Databricks notebook. In this article, I’ll show what I use to read from a Kafka topic that has no schema attached to it. We’ll also dive into how we can render the JSON schema in a human-readable format.

Last week I was working on a Databricks script that needed to produce a Slack message as its final outcome. I lifted some code that used a Slack client that was PIP-installed. Unfortunately, I could not use the package on my cluster. Fortunately, the Slack API is so simple, that you don’t really need a package to post a simple message to a channel. In this blog I’ll show you the simplest way of producing awesome messages in Slack.

When you are training a machine learning image classification model, you often need to resize the images your dataset into smaller ones. When you retrain your model on new data, you resize the images once more. In this blog I’ll share how S3 can be used to cache the resized images.

Today we’ll be looking at sorting and reducing an array of a complex data type. I’m using Databricks to do Spark, but I’m sure the code is compatible. I’ll be using Spark SQL to show the steps. I’ve tried to keep the data as simple as possible. The example should apply to scenarios that are more complex.