Changelog

Changed

• DynamoDB: Added DynamoDB (Preview). Supports AWS credential chain, access keys, auth profiles, and DynamoDB Local. #495, #887
• Hive: Added Apache Hive (Preview). Supports HiveServer2 via Thrift protocol with TCP and HTTP transports. #661
• Impala: Added Apache Impala (Preview). #190
• RavenDB: Added RavenDB (Preview). #252

Fixed

• Results: Fixed copy/paste not preserving null and boolean values between grid cells.

Changed

• Language Server: Now enabled by default. Disable with dbcode.useLanguageServer setting if needed.
• QuestDB: Now generally available (out of preview). #527

Fixed

• BigQuery: Fixed stored procedure DDL script missing parameters.
• BigQuery: Improved introspection to gracefully handle permission-denied errors. Dataset browsing now shows accessible objects (like procedures) even when table access is restricted.
• Elasticsearch: Fixed connection errors with Elasticsearch 7.x and 8.x servers. #890
• MongoDB: Fixed introspection error on system.profile when user lacks listIndexes permission. #863
• PostgreSQL: Fixed connection failure when user lacks SELECT permission on pg_database view.

Changed

• Date/Time Display: Added “Raw” option to date style setting that preserves original database values without formatting or timezone conversion. Maintains full precision (microseconds) as stored in the database.

Changed

• Parameters: Added support for positional parameters ($1, $2, etc.) outside of function bodies. #889
• Authentication Profiles: Extended .pgpass and command auth profiles to CockroachDB, Greenplum, QuestDB, RisingWave, Timescale, and YugabyteDB.
• Oracle: Added support for Oracle 11g when using Instant Client (Thick mode).

Fixed

• Language Server: Fixed false “Unknown column” warnings for stored procedure/function parameters and DECLARE variables.
• Language Server: Fixed JOIN hints not using table aliases in ON clause conditions (e.g., ON a.id = b.user_id now uses both aliases correctly).
• Redis: Fixed authentication failing on Redis versions before 6.0 which don’t support username in AUTH command. #888

Changed

• Authentication Profiles: Added .pgpass file authentication profile for PostgreSQL.
• BigQuery: Added support for cross-project browsing with separate billing project. Users can now browse datasets from external projects (e.g., bigquery-public-data) while using a different project for query billing.
• Athena: Added workgroup support for organizations with IAM policies that restrict query execution to specific workgroups. #880

Fixed

• Connections: Fixed SSL checkbox being locked for known hosts (AWS RDS, etc.) when using SSH tunnels. #883
• Oracle: Fixed connection error on Oracle Free Docker containers where PRODUCT_COMPONENT_VERSION view returns empty results. #885
• DuckDB: Fixed database file remaining locked after disconnecting. #884
• Cloud Providers: Fixed schema caching failing on Windows due to invalid path characters in connection IDs.

Changed

• Microsoft Access: Added support for Microsoft Access databases (.mdb, .accdb). (Preview, Windows only). #546
• Transactions: Added dbcode.transactions.defaultAutoCommit setting to control the default auto-commit mode for new pinned connections.
• Schema Switching: Added automatic schema context detection for Oracle, PostgreSQL, DB2, Databricks, and Snowflake when executing schema switch commands.

Fixed

• SQLite: On large databases (1GB+) using approximate ANALYZE with analysis_limit and skipping table size calculations.

Fixed

• Numeric data types were not being correctly displayed in the grid, preventing charting and other features.
• Refresh and Edit features now work correctly in Notebooks.
• Tables and views in the DB Explorer now show in alphabetical order for MySQL/MariaDB.

Changed

• Added a Show Only Discovered Connections option to the DB Explorer.
• Preliminary support for Oracle, including Instant Client.
• Formatting provider for SQL files and notebook cells now added.
• Added a Copy Name option to the context menu when clicking on a column header.
• Added CURRENT_TIMESTAMP as a valid value for date columns when editing data.
• Added support for organizations in Neon.

Fixed

• Autocomplete correctly escapes identifiers, and does not suggest columns from other tables.
• Tunnel creation and editing can now select files from the .ssh folder when it’s hidden by the OS.

Changed

• Microsoft Access: Added support for Microsoft Access databases (.mdb, .accdb). (Preview, Windows only). #546
• Transactions: Added dbcode.transactions.defaultAutoCommit setting to control the default auto-commit mode for new pinned connections.
• Schema Switching: Added automatic schema context detection for Oracle, PostgreSQL, DB2, Databricks, and Snowflake when executing schema switch commands.

Fixed

• SQLite: On large databases (1GB+) using approximate ANALYZE with analysis_limit and skipping table size calculations.

Changed

• DuckDB: Added extensions support for loading core and community extensions on connect.
• Formatting: Improved SQL formatter dialect mappings for ClickHouse, Databricks, Access, CQL, and Elasticsearch.
• Results: Quick search now highlights matching cells in the grid.

Fixed

• SQLite: Fixed slow connection times on large databases (1GB+) by skipping expensive ANALYZE and table size calculations.
• SQL Server: Fixed backup failing on Docker/local instances with SSL certificate errors by explicitly setting encryption mode.
• Commands: Fixed Quick Open and other commands failing on first attempt when extension hadn’t activated yet. #876

Changed

• TiDB: Added support for TiDB. #872
• Trino: Username and password are now optional.

Fixed

• Zero Config: Fixed symlinked .env files not being detected during connection discovery. #855
• MySQL/MariaDB: Added support for MaxScale proxy with dbfwfilter by auto-detecting and disabling multi-statement queries (or manual connection configuration). #870

Changed

• Language Server: Added dbcode.enableLinting setting to disable SQL linting (syntax and semantic validation). #795

Fixed

• Language Server: Fixed false “Unknown column” errors appearing during initial connection when schema is still loading. #869
• PostgreSQL: Updated bundled pg_dump from 17.6 to 18.1 to support PostgreSQL 18 servers. #871
• Open Table: Now uses workspace default connection when invoked from command palette. #855

Changed

• Elasticsearch: Added support for Elasticsearch (Preview). #323
• QuestDB: Added support for QuestDB (Preview). #527
• ScyllaDB: Added support for ScyllaDB (Preview). #680
• SQLite: Added mod_spatialite extension for spatial SQL capabilities. #867 #576

Fixed

• Cassandra/ScyllaDB: Fixed USE command not updating codelens to reflect the current keyspace.
• Requests: Fixed fetch failing with “Invalid URL protocol” error when proxy is misconfigured or undefined. #868

Changed

• SQL Server/Azure: Added backup/restore support. #195
• MongoDB: Added backup/restore support using mongodump/mongorestore. #195
• MongoDB: Added create and drop database support.
• Trino: Now generally available (out of preview).
• Athena: Now generally available (out of preview).

Fixed

• MongoDB: Fixed introspection error when user does not have collStats permission. #863
• Results: Fixed focus not being set on first cell when opening a table. #840
• Zero Config: Fixed connection IDs to be deterministic, allowing workspace default connections to be shared across developers. #855

Changed

• MongoDB: Added Command authentication provider for retrieving credentials from external secret managers (1Password CLI, Vault, AWS Secrets Manager, etc.) or custom scripts. #496
• Performance: Added experimental background schema worker to build schemas off the main thread. Enable with dbcode.useSchemaWorker setting.

Fixed

• MySQL: Fixed connection timeout on large databases (1000+ tables) by loading database sizes asynchronously. #862
• Authentication Profiles: AWS, auto runs SSO login if needed.