FAQ

Frequently Asked Questions

Is SnoutDB a database server?

No. SnoutDB is a local CLI and embeddable columnar analytics engine. It does not expose a network service.

Does SnoutDB use SQL?

No. Queries use explicit CLI expressions for grouping, filtering, aggregation, sorting, and limiting.

Which inputs are supported?

CSV, JSONL/NDJSON, several log formats, standard input, and .snout files.

What is a .snout file?

It is SnoutDB's versioned, typed, columnar storage format for reusable local datasets.

Can I query a raw file directly?

Yes:

./snout -f data.csv group=region -- count=rows

Create a .snout file when the same data will be queried repeatedly.

Does SnoutDB load large files entirely into memory?

CSV, JSONL, and log inspection/profiling use streaming paths. Some query and aggregation operations materialize typed tables or aggregate state and can therefore require memory proportional to the workload.

Is the file format stable?

It is versioned, but SnoutDB remains pre-v1.0.0. Review release notes before upgrading persisted workflows.

Is Windows supported?

The code is intended to be portable, but current automated validation runs on macOS. Treat unvalidated platforms as experimental.

Where should I ask questions?

Use GitHub Discussions for usage help. Use Issues for reproducible defects and focused feature requests.