Files, Branching & Reactivity
This section covers the parts of AetheriusDB that blur the line between a database, a filesystem, and a version-control workflow. Files appear as tables and vice versa; pipelines react to new data without external schedulers; and the whole database — data and logic — can be branched, merged, and rolled back the way you branch code.
Why it matters
Section titled “Why it matters”These features remove the glue you normally write around a database — the ETL job that loads dropped files, the trigger sprawl that chains transforms, the fragile snapshot process you use to test on real data. The database does those jobs itself, with version-control ergonomics you already understand.
- No loader for files — point the database at a directory and its files are live tables; no import step to schedule.
- No external ETL — pulse tables and cascades transform data source-to-sink on their own, with no triggers or scheduler.
- Branch and roll back like code — experiment on a real copy, then merge or discard; a bad deploy reverts with a checkout.
How fast it is
Section titled “How fast it is”- Branch a live database in milliseconds. A branch is a cheap pointer operation, not a copy — so spinning up a real-data sandbox is instant.
- Reacts the instant a file is closed. A dropped CSV/Parquet becomes queryable the moment its writer finishes, with no polling delay.
- Instant rollback. Reverting is a pointer repoint, not a restore.
In this section
Section titled “In this section”- File-Relational Duality — point the database at a directory and dropped CSV/Parquet files become live, queryable tables the moment a writer closes them.
- Reactive Pipelines — define pulse tables and cascades so the engine transforms data from source to sink automatically, with no triggers or external ETL.
- Database Branching — branch a live database in milliseconds, experiment on real data, then merge or discard.
- Git-Native Source Fabric — tie database logic to Git commits so deploys are a manifest push and rollbacks are a checkout.
Each feature page carries an availability badge (Stable / Opt-in / Beta / Roadmap) so you always know what is production-ready versus still evolving.