Docs / Import, export and backups

Import, export and backups

Result and table exports, CSV, JSON and SQL imports, and complete database backups with optional S3 upload. Desktop only.

All transfer features are desktop-only. Result and table export, import menus, connection archives and database backup are absent on iOS and Android, and the native backend rejects those entry points.

Export

One dialog serves table data, query results and table context menus. Choose columns and a format: CSV, JSON, SQL or genuine Excel .xlsx, optionally gzip-compressed. Export the current page or all matching rows with applied filters. There is no row cap: Go streams rows to a temporary file with bounded memory, progress and cancellation.

SQL is available for named SQL tables and offers CREATE structure, DROP statements and INSERT data. Views support CREATE and DROP only. MongoDB and Redis omit those controls.
CSV escapes quotes and newlines and neutralizes text that spreadsheets could interpret as formulas. Large SQL integers are preserved as strings.
Excel starts another worksheet at its row limit and rejects cells longer than 32,767 characters instead of truncating.
Multi-table exports produce one file per table with independent column selections.
Export files expire after one hour; exports can run for up to 24 hours.

Import

The desktop table context menu supports CSV, JSON and SQL dumps where applicable. QueryBench prepares regular SQL statements and opens them for review before execution, so nothing runs without your confirmation. Edit mode is enforced.

Complete database backup

Back Up Database in the Tools menu creates the selected database's full native backup and streams it into a single local gzip file. The destination is Local file only by default, or Local file and S3 upload. Choose the local backup folder with the native folder picker; the default is the backups folder in the application data directory. Partial files are removed when a backup is cancelled or fails, and completed local files are always preserved, including on upload error. A saved local backup offers Upload to S3 later without rerunning the backup.

Engine	Backup method	Archive
PostgreSQL	pg_dump plain SQL for the selected database	.sql.gz
MySQL	mysqldump including routines, triggers and events	.sql.gz
SQLite	Consistent VACUUM INTO snapshot	.sqlite.gz
DuckDB	Native EXPORT DATABASE with schema and data	.tar.gz

With S3 upload selected, the settings support region, bucket, key prefix, an optional compatible endpoint, AWS profile or default credentials, or access keys with an optional session token. Credentials are remembered per connection in the OS keychain and are never returned to the interface or written to JSON. HTTPS is required for non-loopback custom endpoints. Retry reuses the completed local archive without rerunning the backup.

MongoDB, Redis and ClickHouse do not currently claim full backup support. A table export is never labeled a complete backup.