Learning Spark

A local dev environment to follow the examples in "Learning Spark" 2nd edition by Damji et al.

The dev environment provides PySpark and Jupyter installations to work with the code examples. The Docker Compose file provides PostgreSQL and Kafka (with the corresponding Web-UIs) to try out connecting to them from Spark.

Requirements

• asdf (with the Python plugin)
• Python >=3.11
• Java >=11
• Docker >= 24.0.2
• Poetry >= 1.5.1

To connect to PostgreSQL, the corresponding JDBC driver must be downloaded from https://jdbc.postgresql.org/download/, renamed to posgresql.jar, and placed to the jars directory.

Setup

asdf local python 3.11.2  # or whatever 3.11.x version you have installed
POETRY_VIRTUALENVS_PREFER_ACTIVE_PYTHON=true poetry install

Usage

PySpark and Jupyter

To start Jupyter Lab run

poetry run jupyter-lab

Then open a notebook in the notebooks directory or create your own.

To stop Jupyter Lab press Ctrl-C twice.

To start PySpark shell run

poetry run pyspark

or start a terminal session inside Jupyter Lab and run pyspark.

PostgreSQL and Kafka

To bring up PostgreSQL and Kafka run

docker compose up --detach

See docker-compose.yml for the corresponding ports and credentials of the services.

To stop PostgreSQL and Kafka run

docker compose down --remove-orphans --volumes