DBine v0.1.9

DBine v0.1.9 installers (macOS Apple Silicon and Linux only; Windows and macOS Intel get these changes in 0.1.10).

New

• Rename with impact: Rename… in the explorer changes the name of a table, view, routine, column, index or schema and, in the same script, rewrites the views, procedures, functions and triggers that use it. Before running, it shows what the engine updates by itself, what gets rewritten and what needs a manual check (dynamic SQL, unreadable code), together with the full script. It runs in a transaction where the engine allows it. It is available in every engine that can rename something; each one's limits are in docs/engine-support.md.
• Modify a table: Modify… opens the designer on an existing table and builds the engine's ALTER. It keeps what the designer doesn't show (CHECKs, index options, the order of the key columns) and recreates the views and triggers that depend on the table. Renaming a column there goes through the impact review; on production connections it asks you to type the table name before running.
• Per-query history: the History bar follows the active tab, like a timeline: versions of the saved query with differences and restore, its runs and, in project files, its git commits.
• Navigation in the editor: Cmd/Ctrl+click on a table, view or routine opens its structure or definition, and Show in explorer locates it in the tree. Tables and columns that don't exist are flagged before running.
• Query parameters: :name and ? are asked for when running, and the last value is remembered.
• Snippets per engine (for example, sel + Tab) and a right-click menu in the editor.
• Selection totals: selecting cells in the grid shows count, sum, average, minimum and maximum.
• Scheduled tasks: scripts, exports, schema comparison, backups, Document the database and Send an email (SMTP) that run with DBine closed, through the system scheduler. With notifications per task and a run history. Anything that changes data is approved explicitly.
• Code quality: per-engine rules in the editor and View problems.
• Document the database: a data dictionary in HTML or Markdown, with a diagram, estimated rows and comments on views and routines. Estimated rows and comments come from the engine's metadata, without reading tables or using quota on cloud engines.
• Design query: a visual query builder.
• Copy a subset of data, with masking.
• Optimize query: rewrites, suggested indexes, AI alternatives and a measured comparison. AI alternatives are validated against the database's estimated plan before they are shown.
• Health check of a database, in every engine, with its own checks in SQL Server, the PostgreSQL family, the MySQL family, Oracle, SAP HANA, Firebird, ClickHouse, Snowflake, BigQuery, Databricks and ODBC profiles.
• Search in database: object names, view and routine code, and column names (with their table and type).
• Generate test data for a table.
• Database properties and advanced options when creating a database, in tabs and per engine, with a script preview.
• JSON tree view of results, with editing, and Add row / Add document in the Data tab and in the grid.
• New brand: the logotype with the halo.

Improvements

• The connection color shows as a stripe on the edge of the row, and the dot only indicates the state (green connected, red disconnected).

Fixes

• Schema comparison is no longer cancelled by explorer reads, and SQL Server reconnects.
• Connection rows without a color line up with those that have one.
• Dragging tables into the query builder works on macOS.
• SQL Server properties: long file names no longer overflow the dialog, and the "ANSI and security options" tab is translated.
• The editor no longer flags the columns of an aliased subquery as unknown.

Already available

• Run a query on several databases at once: you pick one or more databases of a connection, and the results are joined with a column that shows each row's database. It arrived in 0.1.4. See docs/multi-database-queries.md.