cube-js/cube · error · Error

Redshift server does not support more than 32767 parameters,

Error message

Redshift server does not support more than 32767 parameters, but ${length} passed

What it means

Redshift's PostgreSQL-wire implementation breaks with more than 32767 bind parameters ('there is no parameter $-32768'), a server-side bug. RedshiftDriver.checkValuesLimit proactively throws a clear error when a query would send 32768+ parameter values.

Source

Thrown at packages/cubejs-redshift-driver/src/RedshiftDriver.ts:294

  protected getInitialConfiguration(
    dataSource: string,
    preAggregations?: boolean,
  ): Partial<RedshiftDriverConfiguration> {
    return {
      // @todo It's not possible to support UNLOAD in readOnly mode, because we need column types (CREATE TABLE?)
      readOnly: false,
      exportBucket: this.getExportBucket(dataSource, preAggregations),
    };
  }

  protected static checkValuesLimit(values?: unknown[]) {
    // Redshift server is not exactly compatible with PostgreSQL protocol
    // And breaks after 32767 parameter values with `there is no parameter $-32768`
    // This is a bug/misbehaviour on server side, nothing we can do besides generate a more meaningful error
    const length = (values?.length ?? 0);
    if (length >= 32768) {
      throw new Error(`Redshift server does not support more than 32767 parameters, but ${length} passed`);
    }
  }

  public override async createTable(quotedTableName: string, columns: TableColumn[]): Promise<void> {
    if (quotedTableName.length > 127) {
      throw new Error('Redshift can not work with table names longer than 127 symbols. ' +
        `Consider using the 'sqlAlias' attribute in your cube definition for ${quotedTableName}.`);
    }

    // we can not call super.createTable(quotedTableName, columns)
    // because Postgres has 63 length check. So pasting the code from the base driver
    const createTableSql = this.createTableSql(quotedTableName, columns);
    await this.query(createTableSql, []).catch(e => {
      e.message = `Error during create table: ${createTableSql}: ${e.message}`;
      throw e;
    });
  }

  /**

View on GitHub (pinned to 7d981676b3)

Solutions

Chunk the upload/insert into batches under 32767 parameters (rows * columns per batch)
Upgrade cubejs-redshift-driver — newer versions chunk large uploads automatically
Use a COPY-based load path (S3 + COPY) for large data loads instead of parameterized INSERTs
Reduce the number of filtered values in generated queries

Example fix

// before
await driver.uploadTable(table, columns, { rows: allRows }); // 100k rows
// after
for (const chunk of chunk(allRows, 5000)) {
  await driver.uploadTable(table, columns, { rows: chunk });
}

Defensive patterns

Strategy: validation

Validate before calling

if ((values?.length ?? 0) >= 32768) throw new Error('Chunk query values below 32767 for Redshift');

Type guard

function withinParamLimit(values) { return Array.isArray(values) && values.length < 32768; }

Try / catch

try { await driver.uploadTable(table, columns, { rows }); } catch (e) { if (e.message.includes('more than 32767 parameters')) { return uploadInChunks(table, columns, rows); } throw e; }

Prevention

Chunk bulk inserts so rows * columns stays under 32767
Use S3 COPY for large loads instead of parameterized inserts
Limit IN-clause sizes in generated queries
Keep the redshift driver updated for automatic chunking

When it happens

Trigger: Executing a query (query/checkValuesLimit called from executeMethods like uploadTable bulk inserts or large IN filters) with a values array of length >= 32768.

Common situations: Uploading a pre-aggregation table with tens of thousands of rows in a single multi-row INSERT; filtering on a dimension with a very large member list (huge IN clause); bulk load of large dimension tables.

Related errors

AI-assisted analysis of cube-js/cube@7d981676b3 (2026-09-02). Data as JSON: /api/errors/4e40fc76fb954fb7. Report an issue: GitHub.