Changelog

Changed

• Scratch Files: Added “none” rotation option for persistent scratch files without timestamp.
• Language Server: Added Clickhouse support. #795
• PostgreSQL: Added support for external tables (FDW). #800
• Filters: Can now be edited/deleted from the explorer, and maintain sync with the results grid.

Fixed

• Export: Fixed grid data not visible in webpage/notebook exports. #799
• Filters: Fixed filters duplicating when saved and not updating across panels and notebooks. #790
• Hover: Fixed table column preview formatting with improved layout, smart column hiding, and better alignment. #797
• PostgreSQL: PostGIS geometry columns now display as readable EWKT format instead of raw hex in result grids.
• Language Server: Fixed case-insensitive column name matching in both editor and grid WHERE filters. #792 #773
• Language Server: Fixed partial keyword matching in autocomplete suggestions. #792

Fixed

• ClickHouse: Fixed infinity and NaN values displaying as null in query results. #794
• Language Server: Fixed diagnostics incorrectly showing for unsupported dialects like ClickHouse. #795
• Editor: Fixed “No SQL to execute” error when executing queries in files without assigned connections. #798
• Results: Fixed initial loading performance. #789
• PostgreSQL: Fixed table row counts showing stale data by using pg_stat_all_tables when available. #791

Fixed

• Results: Fixed tooltip sizing and formatting for JSON content.
• Results: Fixed json cell editor saving JSON with whitespace instead of condensed format.
• SQLite: Fixed attached databases showing main database’s tables instead of their own tables. #784
• Filters: Fixed WHERE clause filters not including schema qualification, causing “table not found” errors. #787
• Connection: Fixed USE statement not triggering database context switch after parser update.

Changed

• Language Server: Included SQLite support.
• Connections: Added file-based post-connection SQL support. #784

Fixed

• ERD: Fixed incorrect cardinality display for junction tables with composite primary keys. #785
• SQLite: Fixed ATTACH DATABASE and DETACH DATABASE statements not being recognized to refresh the explorer. #784
• SQLite: Fixed introspection error when database has read-only file permissions. #783

Changed

• Language Server: Included MySQL and Trino support.
• Filters: Table filters that are saved can now be loaded directly from the explorer.

Fixed

• ClickHouse: Fixed CREATE FUNCTION statement boundary parsing. #781
• Results: Fixed JSON columns being double-stringified when opening grid selections in editor with JSON format.
• Results: Fixed cell editor treating formatting-only changes as data modifications for JSON and XML content.
• Notebooks: Fixed result grid disappearing when clicking refresh or save dropdown menus. #780

Changed

• Results: Added ability to save filters and re use on tables.
• Language Server: Added PL/SQL and T-SQL dialect support to the experimental language server for autocomplete and syntax validation.

Fixed

• Monaco Editor: Fixed theme initialization crash when VS Code themes contain non-string color values.
• Language Server: Disable diagnostics for unsupported dialects to prevent false positives. #772
• Themes: Adjusted some colors from the VSCode theme css provided. #776

Changed

• Results: Added SQL based where filter to results. #745
• Files: Added custom editor support for data and database files (CSV, XLSX, Parquet, Avro, SQLite, DuckDB) - users can now use “Open With…” to open files in DBCode. #764

Fixed

• Autocomplete: No longer suggests items immediately after typing a comma - now requires at least one character to be typed for context-aware suggestions. #765
• Editor: Fixed DECLARE block parsing. #772

Changed

• Editor: Opening SQL now reuses existing untitled files for the same connection instead of creating multiple tabs.
• Editor: Added scratch files for persistent SQL files organized by connection with configurable rotation (hourly/daily/weekly/monthly) and automatic cleanup.
• Editor: Reworked statement boundary parsing to better handle functions, procedures, and other complex SQL constructs. #768 #767
• Connection Roles: Added new statement types for transactions, maintenance, grant/revoke, set/use.
• Explorer: Added table search to right click context menu in explorer, supports multiple table selection.
• Oracle: Added “Include Oracle Maintained Schemas” option to show system schemas (SYS, SYSTEM, etc.). #769
• Redis: Added support for executing statements in the editor.
• Results: Added “Open” to context menu to open data directly in editor. #630

Fixed

• Parameters: Fixed parameter detection incorrectly interpreting string concatenation operator (||) as parameter definitions. #770
• Redis: Follows pagination correctly when loading keyspaces.
• Results: CSV and Excel exports now respect the grid’s column order, including user-reordered columns.
• Table Editing: Not correctly handling tables with columns already set to null.

Changed

• Snowflake: Added support for dynamic tables, hybrid tables, transient tables, external tables and materialized views. #113

Fixed

• All databases: Added more sql keywords to the escape list. #763
• MySQL: Fixed timestamp values in SQL INSERT exports including timezone suffix that MySQL doesn’t accept. #761
• Oracle: Fixed schema detection failing when connecting with system accounts (SYSTEM, SYS, etc.). #760
• SQLite/D1: Fixed introspection error when database metadata references non-existent tables. #667
• Snowflake: Warehouse and Security Role now fetch from the server. #762

Changed

• MongoDB: Added authentication database (authSource) option to connection settings. #752

Fixed

• RisingWave: Fixed connection errors with RisingWave versions below 2.0 due to missing information_schema support. #753
• Results: Expansion now handles more complex nested data, along with correctly displaying row numbers and background stripes. #710
• Bigquery: Fixed JSON data types being quoted. #710

Changed

• ERD: Can now choose to display multiple schemas in the diagram. #744
• Import: You can now create a new table when importing. #529
• Import: Now supports importing from parquet and avro files.

Fixed

• ERD: Filtering to start a specific table was not showing tables that did not contain that table’s name.

Changed

• File Formats: CSV, Parquet, and Excel files now save changes back to disk after editing. #751
• Explorer: Added create capability for database objects (triggers, views, procedures, functions, indexes) with dialect-specific SQL across all supported databases. #67
• Results: Added Parquet export format. #730
• Results: Added option to expand a JSON column into unique keys or rows if an array. #710
• Notebooks: Added python language support. #621

Fixed

• Results: Fixed “Export - All” to automatically fetch all rows when row limit is reached. #750

Changed

• MCP/Copilot: Added separate dbcode-execute-dml and dbcode-execute-ddl tools for granular permission control. #739
• MCP: Breaking Change - Legacy auth modes (Bearer, Querystring) and depreciated SSE transport removed. OAuth provides automatic discovert and secure authentication with approval dialogs for MCP clients, and no need to set query strings, tokens etc. URL has also been changed to http://localhost:PORT/mcp.
• MongoDB: Added create, drop, and rename capabilities to MongoDB collections, matching functionality available for SQL database tables. #66

Fixed

• Connections: Fixed refresh not updating schema metadata when database structure changes (table field types, new databases, etc.). #747
• Azure/SQL Server: Stored procedures and functions were not scriptable. #743

Changed

• Sign in: Added web based sign in option including support for Google. #742

Fixed

• Results: Fixed type collision issue when using data formatters that caused some values to not display.

Changed

• Backup/Restore: Added backup and restore functionality for PostgreSQL MySQL, SQLite, and DuckDB to start to the right click context menu of their databases (uses native tools). #195