SQL Analytics Studio
Run SQL queries on loaded datasets and export query results.
A clear question. A clearer answer.
SQL guide & examples
DuckDB supports familiar analytical SQL, with dialect differences. Examples use mounted names and never execute automatically.
Mount a table to prepare an example.
Derived columns in SELECT are result columns. For a reusable query, use a view. ALTER TABLE features vary by engine; adding a generated column after creation is not supported here. Date text formats must be explicit; TRY_CAST/TRY_STRPTIME return NULL when conversion fails. Double quotes identify tables/columns; single quotes represent text values.
History & saved queries
Keep up to 30 successful queries and 50 named queries in this browser. SQL text may contain sensitive literals. Tables and results are not saved with queries; remount files after refresh. Loading never executes SQL.
Saved queries
Recent successful queries
Drop CSV, JSON or Parquet files here to mount tables
Files mount as separate tables. Existing tables are kept; repeated names get a numeric suffix.
A place for better questions.
Mount a local file to explore its tables with SQL.
| Result |
|---|
| Execute a query to see results here. |
Follow a question.
Find the connection.
A browser SQL workspace for querying local tables with DuckDB. Inspect a file, compare groups, or combine tables using explicit queries.
How to query a dataset
- Mount your files. Wait for the engine to become ready, then mount CSV, JSON or Parquet files. Each file becomes a temporary table.
- Inspect the schema. Check table names, column names and detected types. Quote identifiers with double quotes when needed.
- Ask your question. Start with SELECT, add WHERE or GROUP BY, or insert a schema-based example from the SQL guide. Inserting does not execute it. Run the query when ready; DuckDB syntax can differ from other databases.
- Keep your result. Browse results in pages of 100 rows and export the complete successful result as CSV. Save a named query or revisit successful history. Tables must be remounted after refresh.
Which SQL does SQL Studio use?
SQL Studio uses DuckDB SQL. Common SELECT queries, joins, grouping and window functions are available, but database dialects differ. MySQL or SQL Server syntax may need adjustment. The SQL guide provides examples using your mounted tables; inserting an example never runs it automatically.
Which files can I query?
Mount CSV, JSON or Parquet files as temporary tables and join them in a query. Inspect detected names and types in the schema. Double quotes identify tables or columns, including names with spaces; single quotes represent text values.
How do I add a derived column or parse dates?
Use a SELECT expression to add a result column without changing the imported table. The guide includes YEAR(TRY_CAST(column AS DATE)) for date or ISO text values, and TRY_STRPTIME for explicit text date formats. Failed try-conversions become NULL. Adding generated columns after table creation is not supported by this engine.
What is saved, and does my data leave the browser?
Files are processed by browser DuckDB; the bundled engine loads from this site. SQL that accesses remote URLs or installs extensions can make additional network requests. Query text, up to 30 recent successful queries and 50 named queries can be stored in this browser. SQL can contain sensitive literal values. Storage may be unavailable, cleared or full; tables and query results are not persisted. Remount files after refresh and download useful results.
Does exporting include every page, and what are the limits?
Results display 100 rows per page; CSV export includes the complete most recent successful query result, regardless of the page shown. Empty results have no CSV export. Results stay in Arrow format in browser memory; only the visible page is converted for display. CSV export converts rows in chunks, but the download still needs memory. Pagination does not limit engine memory use. This is a temporary analytical workspace, not a persistent database server.