Enter SQL in the SQL editor tab and press ⌘↵, or select Run SQL. Multiple statements are supported. Each statement produces its own result record, including returned columns and rows or the affected row count.
The statement splitter recognizes quoted strings, quoted identifiers, and SQL comments. It is not a complete dialect parser; use engine-native tools for procedural scripts, dollar-quoted bodies, and client delimiter commands.
Read results appear in a grid. Write statements report affected rows. A failed run clears the current results and displays the database error.
Read results are bounded to 10,000 rows / 16 MiB. Use table export when you need all rows. The toolbar provides direct controls for formatting, history, favorites, explain, transaction execution, charts, export, and copying Markdown. Actions that require SQL or results are disabled until those are available.
Executed queries are saved locally with execution time and any error. Open History to inspect recent statements and load one back into the editor. History is capped to prevent unlimited file growth.
Open Favorites, provide a name, and save the current editor text. Selecting a favorite loads its SQL. Removing a favorite deletes it from local metadata only.
When a result contains columns and rows, open the chart action to create a bar, line, or pie view. Choose columns with compatible label and numeric data. Charts are exploratory and do not modify the database.
- Use transactions for related changes when supported by your script.
- Add a
WHEREclause before runningUPDATEorDELETE. - Use a database read-only account for investigation.
- Start with a
SELECTusing the same predicate as a planned write. - Keep backups for schema or bulk changes.