dbx VS nutter

Pyspark is the python API for spark. You can write code in a notebook on databricks and run it on a cluster or you can write code in an IDE and run it using dbx through the dbx execute command. If you’re more familiar with Pandas API, you can use Koalas which is a pandas API on Spark

Hey, Databricks person here. Check out DBX for a template on how to do unit and integration tests: https://github.com/databrickslabs/dbx

Approximately 1 year ago I've released the first version of dbx - a CLI tool for simple and efficient development and deployment of Databricks jobs.

From my experience OO gives you much more flexibility in designing your pipeline but you're risking to make the project way more complicated. The worst example I have seen is the Nutter library (https://github.com/microsoft/nutter), which uses endless classes that are all nested in each other. I once had a bug when using it, and it was a huge pain in the ass to understand what's going on when the code is executed. It is a very good example of what can go wrong when you're overusing OO. However, in one project, I carefully created few classes, just out of curiosity, and I was very impressed how it helped me to organize/structure my code. A functions hase a clear dedicated use, but a good class is like a Swiss army knife with an solid set of functionalities. If you know how to use it in a smart way, you are likely to increase the quality of your code, but the contrary is also very likely, especially when the team members are not ready for it.

- https://github.com/microsoft/nutter