Version Control

csvs dataset SHOULD be version controlled.
credentials and other sensitive data SHOULD be stored separately from the dataset, e.g. in .git/config or in another csvs dataset under access control.

Concurrency

csvs is designed for single-writer access. Only one process should write to a dataset directory at a time. Running panrec and evenor against the same dataset simultaneously may corrupt tablet files, because there is no file-locking mechanism.

For multi-device access, use git-based sync: each device commits its changes locally and merges via git pull/git push. Git's merge semantics handle concurrent edits at the line level, and evenor's resolve operation performs three-way merge when conflicts arise.

If a client builds the schema once and reuses it across operations (see the schema option in csvs-js or with_schema() in csvs-rs), note that the cached schema becomes stale after any operation that modifies _-_.csv. Rebuild the schema after schema-modifying writes.

To learn more about csvs, see other User Guides and the Requirements.