You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Group several SQLite files under one connection #2536
Each SQLite file is its own connection. A project holding several, which is normal when a schema is split by concern, fills the connection list with entries that belong together. A file attached with ATTACH in the editor can be queried as alias.table, but databases/sqlite.mdx (Browsing) says it "never appears in the sidebar": the object list reads the main database only. The attachment is gone after a reconnect, so it has to be typed again.
DuckDB is halfway there: an attached file joins the sidebar as a sibling database after a refresh and Cmd+K switches to it, but DuckDB does not persist attachments, so a reconnect starts again from the one file (databases/duckdb.mdx).
Proposed solution
Several file paths on one SQLite connection, each attached under a name the user picks, defaulting to the file name.
Each attached database as its own node in the sidebar, with its tables under it, browsable and editable like the main one.
Cross-database queries work, since ATTACH is what makes them work, and completion knows alias.table.
The attachments are part of the saved connection, so every connect, reconnect and restored tab attaches them again before anything runs.
The same for DuckDB local files and libSQL local files, which use the same ATTACH.
A file that is missing at connect time fails with its path named, and the rest still attach.
Done means a connection saved with app.db, auth.db and audit.db lists three databases in the sidebar after a relaunch, and SELECT … FROM auth.users JOIN audit.events … runs.
Alternatives considered
One connection per file, and ATTACH typed again after every reconnect.
Activity