windmill-labs/windmill · error

BigQuery requires a dataset (schema) name

Error message

BigQuery requires a dataset (schema) name

What it means

BigQuery organizes tables under datasets, and its INFORMATION_SCHEMA queries require the dataset name as the schema qualifier. makeBigQueryForeignKeysQuery therefore refuses to run when schemaName is undefined/empty, because the generated `dataset.INFORMATION_SCHEMA.*` SQL would be invalid or ambiguous. Note makeForeignKeysQuery passes schemaName through as-is (no default), so unlike snowflake/duckdb there is no fallback.

Source

Thrown at frontend/src/lib/components/apps/components/display/dbtable/queries/relationalKeys.ts:316

function makeSnowflakePrimaryKeyQuery(tableName: string, schemaName: string): string {
	return `
SELECT
    constraint_name
FROM
    information_schema.table_constraints
WHERE
    constraint_type = 'PRIMARY KEY'
    AND table_name = '${tableName}'
    AND table_schema = '${schemaName}'
LIMIT 1;
`.trim()
}

// BigQuery queries
function makeBigQueryForeignKeysQuery(tableName: string, schemaName?: string): string {
	if (!schemaName) {
		throw new Error('BigQuery requires a dataset (schema) name')
	}

	return `
SELECT
    tc.constraint_name as fk_constraint_name,
    kcu.column_name as source_column,
    ccu.table_name as target_table,
    ccu.column_name as target_column,
    'NO ACTION' as on_delete,
    'NO ACTION' as on_update
FROM
    \`${schemaName}.INFORMATION_SCHEMA.TABLE_CONSTRAINTS\` tc
    JOIN \`${schemaName}.INFORMATION_SCHEMA.KEY_COLUMN_USAGE\` kcu
        ON tc.constraint_name = kcu.constraint_name
    JOIN \`${schemaName}.INFORMATION_SCHEMA.CONSTRAINT_COLUMN_USAGE\` ccu
        ON tc.constraint_name = ccu.constraint_name
WHERE
    tc.constraint_type = 'FOREIGN KEY'

View on GitHub (pinned to e474e8803c)

Solutions

1. Always pass the BigQuery dataset as schemaName when calling makeForeignKeysQuery.
2. In the UI, require dataset selection before enabling the relations panel for bigquery dbType.
3. Derive the dataset from the fully-qualified table reference (project.dataset.table) if only the table was given.
4. Use a sensible default dataset in the app config if the app always targets one dataset.

Example fix

// before
const q = makeForeignKeysQuery('bigquery', 'my_table', undefined)
// after
const q = makeForeignKeysQuery('bigquery', 'my_table', 'my_dataset') // project.my_dataset.my_table

if (dbType === 'bigquery' && !schemaName) {
  throw new Error('Select a BigQuery dataset before exploring relations')
}
const q = makeForeignKeysQuery(dbType, tableName, schemaName)

const hasBigQueryDataset = (s: unknown): s is string =>
  typeof s === 'string' && s.trim().length > 0

try {
  const q = makeForeignKeysQuery('bigquery', tableName, schemaName)
} catch (e) {
  if (e instanceof Error && e.message.includes('dataset (schema) name')) {
    showToast('BigQuery requires a dataset — select one first')
    return null
  }
  throw e
}

When it happens

Trigger: Requesting foreign keys for a BigQuery table while the schema/dataset field is empty — e.g. the user did not pick a dataset in the table editor, the schema came from an unfilled app variable, or the table reference was entered without a dataset.

Common situations: BigQuery resource where tables are browsed without selecting a dataset first; app config storing only the table name; migrations or UI changes that stopped propagating schemaName.

Related errors

• Column ${column.field} is not nullable and has no default va
• Unsupported database type: ${dbType}
• Invalid language: ${language}
• Unsupported database type:${dbType}
• Unsupported database type

AI-assisted analysis of windmill-labs/windmill@e474e8803c (2026-09-03). Data as JSON: /api/errors/f886662fcac08261. Report an issue: GitHub.