`bigquery_attach`¶

Enumerates the native tables in one BigQuery dataset and creates local DuckDB views backed by bigquery_scan.

Compatibility helper

Prefer ATTACH ... (TYPE bigquery) for catalog, transaction, DDL, and write integration.

Signature¶

CALL bigquery_attach(
    'PROJECT.DATASET',
    overwrite := false
);

Parameters¶

Parameter	Type	Default	Description
`dataset`	`VARCHAR`	required	Project and dataset identifier.
`overwrite`	`BOOLEAN`	`false`	Replace colliding local views.

Example¶

The current implementation emits no row. The output below comes from querying one of the local views it creates. Because the view name is the fully qualified BigQuery table identifier, quote it as one DuckDB identifier.

-- Create local compatibility views for the dataset's native tables.
CALL bigquery_attach(
      'my-gcp-project.quacking_dataset',
      overwrite := false
  );

-- Query one of the generated local views.
SELECT
      i,
      s
  FROM "my-gcp-project.quacking_dataset.duck_tbl"
  ORDER BY i;
┌───────┬────────────────┐
│   i   │       s        │
│ int64 │    varchar     │
├───────┼────────────────┤
│    12 │ quack 🦆       │
│    13 │ quack quack 🦆 │
└───────┴────────────────┘

Result and Behavior¶

The function declares one Success BOOLEAN result column. The current implementation performs the view-creation side effects but emits no row, so invoke it with CALL.

Reading the created views uses bigquery_scan and therefore inherits its Storage Read permissions, cost, and native-table limitations. The function uses the direct project's default credentials and endpoints; it does not reuse an attached catalog.

Typical errors include an identifier other than project.dataset, failed dataset discovery, missing permissions, and a local view-name collision when overwrite=false.

See ATTACH for the recommended attachment workflow.

bigquery_attach¶