This repository contains scripts to attempt to profile SQLite database file sizes when Write Ahead Logging (WAL) mode is enabled.

This isn't the most elegant code and went through some iterations, but should hopefully be "good enough" for its intended purpose. For example, it was originally intended to just produce ".csv" files of the results and interpret them in LibreOffice/Excel. However, midway through the investigation it was decided to experiment with matplotlib instead. It is likely evident that this was cobbled together as an afterthought.

This repository contains two executable scripts:

• main.py: Starts a process to monitor file sizes then begins running SQLite test scenarios. The results are saved both as .csv files and pickled data (.pickled).
• plotter.py Reads in the .pickled files and produces graphs using matplotlib.

The results of running these scripts are available in the results directory. A blog post discussing the results can be found here.

The following is an example of the commands required to run and plot a single scenario. Without an argument, main.py will start running all the test scenarios.

./main.py --scenario 0
./plotter.py --show_plot results/${new_timestamped_folder}/results_scenario_0.pickled