rosbag2csv

Simple utility to export tables from a sqlite3 database (generated from a ROS bag) to CSV.

This repository contains a tiny CLI helper (read_bag.py) that lists tables from a sqlite3 database and exports a selected table to a CSV file.

Features

• List tables in a sqlite3 database
• Export a chosen table to <table>.csv (writes header row)

Requirements

• Python 3.7+
• Standard library only (no external dependencies)

The script uses the built-in sqlite3 module, argparse, and csv.

Usage

Place your sqlite3 database file (for example, a ROS bag-derived database) somewhere accessible and run the script.

Show available tables:

python3 read_bag.py --db /path/to/your/dbfile.db3 --list

Export a table to CSV (writes TABLE_NAME.csv in the output directory):

python3 read_bag.py --db /path/to/your/dbfile.db3 --table TABLE_NAME --out /path/to/outdir

Notes:

• If --out is omitted the CSV is written to the current directory.
• If --db is omitted the script uses a project-local default path embedded in the script; pass --db to select a different database.

Exit codes

The script uses numeric exit codes to indicate error types. Common values:

• 0: success
• 1: invalid usage (e.g. missing --table and not using --list)
• 2: requested table not found or no tables available
• 3: database file not found
• 4: could not open database
• 5: error checking table existence
• 6: error querying database
• 7: error writing CSV file

Contributing

Small, focused pull requests are welcome. If you want to:

• Add unit tests, please use pytest and add a minimal test that runs quickly.
• Switch to a more ROS-native workflow (for example, using rosbag2 tooling) open an issue first to discuss scope.

TODO

Planned improvements and areas for future work:

• Decode CDR-serialized message payloads before export: many ROS message fields are stored in CDR (Common Data Representation). Adding a deserialization step will produce human-readable fields that can be written to CSV columns rather than raw binary blobs.
• Organize CSV outputs by topic: write CSV files grouped by ROS topic (for example, one CSV per topic or topic-named subdirectories) so data from different topics stays separated and easy to consume.
• Add caching / buffering for large datasets: introduce an in-memory or on-disk cache and batch-writing strategy to improve throughput and reduce memory pressure when exporting large databases.

License

MIT License — see LICENSE (not included). If you want a different license, please open an issue.

Contact

Open issues / pull requests in the project repository.