What causes "Streaming is not supported by the SingleStore Proxy driver" in drizzle-team/drizzle-orm?

Replace iteration with execute()/all() and process the full result array, paginating via LIMIT/OFFSET or keyset pagination for large sets.

drizzle-team/drizzle-orm · error · Error

Streaming is not supported by the SingleStore Proxy driver

Error message

Streaming is not supported by the SingleStore Proxy driver

What it means

Thrown by PreparedQuery.iterator (singlestore-proxy/session.ts:168). The proxy driver issues whole-result requests over the remote callback and has no streaming protocol, so async iteration over result rows is not possible; calling .iterator() (or for-await on the query) is rejected.

Source

Thrown at drizzle-orm/src/singlestore-proxy/session.ts:168

				return returningResponse;
			}

			return data;
		}

		const { rows } = await client(queryString, params, 'all');

		if (customResultMapper) {
			return customResultMapper(rows);
		}

		return rows.map((row) => mapResultRow<T['execute']>(fields!, row, joinsNotNullableMap));
	}

	override iterator(
		_placeholderValues: Record<string, unknown> = {},
	): AsyncGenerator<T['iterator']> {
		throw new Error('Streaming is not supported by the SingleStore Proxy driver');
	}
}

export interface SingleStoreRemoteQueryResultHKT extends SingleStoreQueryResultHKT {
	type: SingleStoreRawQueryResult;
}

export interface SingleStoreRemotePreparedQueryHKT extends SingleStorePreparedQueryHKT {
	type: PreparedQuery<Assume<this['config'], SingleStorePreparedQueryConfig>>;
}

View on GitHub (pinned to b7862528fd)

Solutions

Replace iteration with execute()/all() and process the full result array, paginating via LIMIT/OFFSET or keyset pagination for large sets.
Switch to a streaming-capable driver if true streaming is required.
Cap result size with .limit() and loop in application code over paginated batches.

Example fix

// before (proxy driver)
for await (const row of db.select().from(bigTable)) { handle(row); }
// after
const pageSize = 1000;
let offset = 0;
while (true) {
  const rows = await db.select().from(bigTable).limit(pageSize).offset(offset);
  if (!rows.length) break;
  for (const row of rows) handle(row);
  offset += pageSize;
}

Defensive patterns

Strategy: type-guard

Validate before calling

function driverSupportsStreaming(db) {
  return !/SingleStoreRemoteSession|SingleStoreProxy/.test(db?.session?.constructor?.name ?? '');
}
if (!driverSupportsStreaming(db)) {
  // use pagination instead of iterator()
}

Type guard

function isProxyDriver(db) {
  return /SingleStoreRemoteSession|SingleStoreProxy/.test(db?.session?.constructor?.name ?? '');
}

Prevention

Use execute() + LIMIT/OFFSET pagination instead of iterator() on the proxy driver.
Detect the driver and choose streaming vs pagination accordingly.
Cap page size to bound memory in serverless environments.

When it happens

Trigger: Calling db.select(...).iterator() or `for await (const row of db.select(...))` on a database built with drizzle-orm/singlestore-proxy; using streaming to handle large result sets in an edge/proxy deployment.

Common situations: Porting streaming code from the node mysql2 driver to the proxy driver; memory-conscious large exports that rely on iterator() in a serverless function.

Related errors

AI-assisted analysis of drizzle-team/drizzle-orm@b7862528fd (2026-08-03). Data as JSON: /data/errors/c5f9fe15194aac5c.json. Report an issue: GitHub.