Database

A tool ref to some of its operations, for tools: "query" (a read-only statement), "execute" (a statement that writes) and "schema" (its tables and columns). The database itself in tools gives all three. "upsert" writes one row of one table by its primary key, its parameters that table's columns; bind its table with .bind({ table }).

Create a database, unique by name per user (a taken name throws Conflict with existingId).

A lazy reference to the database called name for the current (or given) user. With createIfMissing it's created on first use with engine; one that exists is used as it is.

In a preview branch of an app whose preview_resources is branch, the branch gets its own copy at its first write and reads production's as it is until then. A Postgres copy is a copy-on-write branch of production's data; a SQLite copy holds the schema and migrations only (migrate's seed fills it), unless branchData: "copy" asks for production's rows too, read at one moment while production keeps serving (up to 200,000 rows and 16 MB). It changes nothing outside a preview.

A lazy reference to a database by id.

Load a database by id now; an unknown id raises noclick.NotFound.

The current user's databases, branches included (or those userFilter covers).

One page of databases; pass nextCursor back as cursor.

Run a read: one select (or with ... select). It runs read-only: a statement that writes throws InvalidRequest; use execute for those. At most 10 MB of rows; 30 seconds. On Postgres, noclick.embed(:q, model) turns a parameter or a quoted text into a halfvec before the statement runs (up to 100 distinct texts a request; not over connectionUrl), so a read (an agent's query tool too) can search by meaning (billed as AI to the database's user).

await kb.query("select id from articles order by embedding <=> noclick.embed(:q, :model) limit 5", { q, model });

Run one statement that writes (several go in a batch). A SQLite database that would pass 10 GB throws QuotaExceeded, whose next step is branching into Postgres.

Run up to 100 statements in one transaction: all apply or none do.

Apply the migrations the database hasn't run, in name order, in one transaction (a failing one applies none). In Node, read your migrations/*.sql files into an object first. dryRun reports what would run and changes nothing. seed is SQL run once, after them and in their transaction, on a preview branch's database that holds none of production's data (a SQLite copy made without its rows, or one only the branch has); production never runs it, and it shows in applied as (seed).

The migrations this database has applied, oldest first.

The database's tables, their columns and row counts.

Put the database back as it was at at (within its history: 30 days on SQLite, 7 on Postgres), in place. wait() resolves to this database, restored. To keep the present, branch at that time instead.

Copy the database into a new one called name, as it is now or as it was at (within its history). engine: "postgres" branches SQLite into Postgres. While a SQLite database is copied its status is branching. wait() resolves to the new Database.

The databases branched from this one.

A postgresql:// URL with a password that stops working after ttlS seconds (default 3600), when any session still open on it ends. SQLite throws NotAvailable. Treat it as a secret.

Export as SQL that recreates its schema and rows in any database of its engine: sqlite3 booking.db < databases/booking-db.sql, or a pg_dump script for psql (NoClick's own noclick schema, owners and grants left out). Exporting a Postgres database wakes it, and its compute is billed as any use is. wait() resolves to an ExportBundle: a zip at url holding the dump.

Delete the database and its data: it goes to the trash, restorable for 30 days, and its connection URLs stop working. One with branches, an app's binding or an agent's tool throws Conflict. Returns its TrashItem.