# Basic Configuration

**A handful of defaults and best practices that keep your OrcaSheets work fast and predictable, especially on larger datasets.**

## Performance basics

If you are working with large datasets:

- **Select only the columns you need**: avoid `SELECT *`, so queries stay fast.
- **Filter early**: narrow by date ranges, partitions, or key subsets before aggregating.
- **Start narrow, then widen**: begin with smaller result sets and expand only when you know the shape you want.

For performance, OrcaSheets may display up to ~1,000 rows at a time in the grid. Calculations and exports still use the **full** dataset.

## Optimisation

Imported files are optimised for fast analysis. Large files may take a moment to optimise, and the status bar shows progress so you know when it is safe to start working.

## Credentials and access

Use least-privilege credentials so you keep blast radius small:

- **Read-only access** for exploration, so you cannot accidentally mutate source data.
- **Scoped access** (specific schemas or tables) wherever the source supports it.
- **Separate credentials per environment** (dev, staging, prod), so a leaked dev key never unlocks production.

## Related pages

- [Connectors](/docs/connectors)
- [Performance Issues](/docs/performance-issues)
- [Access Control & SSO](/docs/access-control-sso)
- [Local-first Security](/docs/local-first-security)