d1Executor

A SqlExecutor over Cloudflare's D1, implementing the asynchronous half of the port that Alchemy's D1 executor implements (@alchemy.run/cloudflare/D1/ApplyMigrations, makeD1MigrationExecutor): query/get for reads, write/batch for writes, and atomic for a unit.

Source: packages/paykit/src/Store/D1.ts Kind: Primitive

A SqlExecutor over Cloudflare’s D1, implementing the asynchronous half of the port that Alchemy’s D1 executor implements (@alchemy.run/cloudflare/D1/ApplyMigrations, makeD1MigrationExecutor): query/get for reads, write/batch for writes, and atomic for a unit.

D1 has no interactive transaction and no isolation between a read and a commit, so atomic collects the unit’s writes and commits them in one batch call — the closest thing the platform offers, and what Alchemy’s own driver does for D1. isolation says "deferred" so a store that needs more can refuse rather than assume.

The synchronous half refuses with an error naming the constraint rather than blocking a request thread that cannot block.

Reaching D1 from a Worker

const sql = Paykit.d1Executor(env.BILLING);
yield* sql.atomic(
  Effect.gen(function* () {
    yield* sql.write("INSERT INTO billing_event (…) VALUES (?, ?, …)", [...]);
    yield* sql.write("INSERT INTO billing_state (…) VALUES (?, ?, ?)", [...]);
  }),
);