Import and Export

Moving Data In and Out

Sooner or later someone hands you a spreadsheet. Sooner or later someone asks for one.

In

Marketing sent leads.csv:

name,email,signed_up
Linus,linus@example.com,2026-03-01
Margaret,margaret@example.com,2026-03-04

Open the import wizard from the command bar (⌘K, "import"), pick the file, and Quarry takes a look. It detects CSV or JSON, the delimiter and the encoding, and guesses types per column: signed_up looks like a date, so it's a date.

Import into an existing table, or let Quarry create one from the guessed types. Creating the table and loading the rows happen in one transaction, so a bad row on line 9,000 doesn't leave you with half a table.

What about duplicates? Pick one:

• Error: a plain INSERT. A conflict stops the import. The safe default.
• Skip duplicates: ON CONFLICT DO NOTHING. Rows already there stay as they are.

Out

Run a query, then ⇧⌘E. You get three formats:

• CSV for the spreadsheet people.
• JSON for the script people.
• SQL INSERT statements, for moving rows to another database.

Export streams. It doesn't build the whole file in memory first, so it works for results much larger than the grid will show you (the grid stops at 100,000 rows).

Whole databases

For a full dump, ⌘K → "Dump Database to SQL…". Postgres and MySQL use pg_dump and mysqldump if you have them; SQLite has its own built in.