Merged
Conversation
- New SQL completion engine using sqlparse for context detection - Suggests tables after FROM/JOIN, columns after WHERE/SELECT, operators after column names - Lazy loading of columns to avoid UI lag during schema indexing - Scrollable autocomplete dropdown with keyboard navigation - Auto-hide when cursor moves away from typing position - Fuzzy matching and alias resolution support 🤖 Generated with [Claude Code](https://claude.com/claude-code)
Previously, an empty string password was treated differently from None, meaning connections with empty passwords would fail instead of prompting the user. Now both None and empty string trigger the password prompt. 🤖 Generated with [Claude Code](https://claude.com/claude-code)
Azure SQL Database does not support cross-database queries or USE statements. This fix: - Adds _get_cursor_for_database() that tries USE first (fast for regular SQL Server), then falls back to creating a new connection with the target database (for Azure SQL) - Stores config and driver on connection object for reconnection - Removes all cross-database reference syntax ([Database].schema.table) - Updates all adapter methods to use the new approach Also adds unit tests verifying no cross-database references are used, and integration tests for Azure SQL Database. 🤖 Generated with [Claude Code](https://claude.com/claude-code)
Reorganizes sql_completion.py into a package with separate modules: - core.py: base classes and utilities - completion.py: main SQLCompletion class - alter_table.py, create_table.py, create_index.py, create_view.py, delete.py, drop.py, insert.py, truncate.py, update.py: statement handlers Also introduces QueryTextArea widget that wraps TextArea with SQL autocomplete functionality, simplifying the integration. 🤖 Generated with [Claude Code](https://claude.com/claude-code)
- Return empty completions when there's no SQL content yet (just whitespace). This prevents showing irrelevant keywords when pressing space/enter on an empty query. - Add handlers for IN ( and EXISTS ( patterns to suggest SELECT for subqueries. 🤖 Generated with [Claude Code](https://claude.com/claude-code)
For Azure SQL Database, clicking a table from a different database now correctly executes the query against that database: - build_select_query no longer includes [database] prefix (breaks Azure) - Added _query_target_database to track which database to query - Query execution uses apply_database_override when target differs - Target database is cleared after use to not affect user-typed queries 🤖 Generated with [Claude Code](https://claude.com/claude-code)
…database When connected without a default database (multiple databases visible), autocomplete now suggests fully qualified names like [db].[schema].[table] instead of just table names. This allows queries to work correctly. Uses adapter.quote_identifier() for proper quoting (brackets for SQL Server). 🤖 Generated with [Claude Code](https://claude.com/claude-code)
Idle Scheduler:
- Add idle scheduler inspired by browser's requestIdleCallback API
- Execute background work during user idle periods (500ms threshold)
- Queue schema loading (tables/views/procedures) as idle jobs
- Add --debug-idle-scheduler flag to show scheduler status bar
- Preload columns for tables found in query editor during idle
Shared Cache:
- Tree and autocomplete now share _db_object_cache for tables/views/procedures
- Expanding tree nodes populates autocomplete cache and vice versa
- Refresh ('f') clears both object cache and column cache
Azure SQL:
- Add separate AzureSQLAdapter with required database field
- Simplify MSSQL adapter by removing Azure-specific fallback logic
Autocomplete UX:
- Add *, DISTINCT, TOP, ALL to SELECT clause completions
- Remove table names from SELECT suggestions (belong after FROM)
- Dropdown stays visible when window loses focus
- Dropdown constrained to stay within bounds (no right-edge cutoff)
- Enter now hides the autocomplete dropdown and inserts a newline - Only Tab accepts autocomplete suggestions - Added suppress flag to prevent autocomplete from re-triggering after Enter dismisses it (newline text change was re-showing dropdown)
Introduces _active_database separate from current_config.database to preserve tree structure while allowing users to set a preferred database. - Added get/set_database_preference() to config.py for persistence - On connect: load cached preference into _active_database - On 'u' press: save preference and reload schema for new database - On disconnect: clear _active_database - Star display in tree now uses _active_database - Autocomplete schema loading now respects _active_database The connection config (and tree structure) remains unchanged - only the user's runtime preference is cached in settings.json.
- Add system_databases property to DatabaseAdapter base class - Define system databases for each adapter: - MSSQL: master, tempdb, model, msdb - PostgreSQL: template0, template1 - MySQL/MariaDB: mysql, information_schema, performance_schema, sys - ClickHouse: system, information_schema - Snowflake: SNOWFLAKE - Update autocomplete to filter using adapter.system_databases - ESC key now closes autocomplete AND exits to normal mode - Add unit tests for system_databases property
Add priority=True to escape bindings on all modal screens to ensure ESC key is handled correctly when other bindings are active.
Flatten multi-line error messages to single line since DataTable cells only display one line. Replaces textwrap with regex whitespace collapse.
- Add supports_cross_database_queries property to DatabaseAdapter - Returns True for MSSQL, MySQL, ClickHouse, Snowflake (support USE) - Returns False for PostgreSQL, CockroachDB, D1 (require reconnection) - Add switch_database() method to ConnectionSession for reconnecting to a different database while reusing SSH tunnel - Auto-switch database when expanding database nodes or their contents - For cross-db adapters: sets active database (USE approach) - For non-cross-db adapters: reconnects to the database - Falls back to reconnection if USE fails (e.g., Azure SQL) - Remove 'u' keybinding - Enter on database node now sets it active - Apply active database to query execution context - Add helper functions in providers.py for database requirement checks - Update tests with new mock attributes
Azure SQL is now handled by the MSSQL adapter with automatic fallback: - If USE statement fails, automatically reconnect to target database - Retry the operation after reconnection - Show original error only if reconnection also fails This makes Azure SQL work transparently without a separate adapter.
Previously, typing a semicolon after a complete statement would show irrelevant suggestions (tables or keywords). Now the autocomplete popup hides when the cursor is immediately after a semicolon.
- Only prompt for password when password is None (not set), not when it's an empty string (explicitly set to empty) - Some databases allow passwordless connections, so "" should be valid - Remove Azure SQL adapter tests (adapter was removed)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This update introduces a sophisticated SQL completion engine with context-aware DML/DDL suggestions, leveraging schema caching and lazy loading table metadata for autocompletions, and a new idle scheduler for background jobs.
It also adds cross-database query handling with capability detection, adapter improvements for various databases.
If we have not selected a default database, It adds a UI indicator of what database we are connected to.
We will automatically "use" a database when we expand on it using 'Enter' replacing 'u' to use.
Table expansion and querying is now more robust for databases that does not support cross-db queries.
This will fix postgres and D1 tables expanding as (empty) when not having selected a default database, as well as querying errors when connected to an Azure SQL database. For databases such as these, will automatically default to reconnecting behind the scenes, instead of "use," while giving the user the experience of still being connected to the server.
Fixes #43
Highly likely fixes #4