> ## Documentation Index > Fetch the complete documentation index at: https://docs.summation.com/llms.txt > Use this file to discover all available pages before exploring further. # Attach connection datasets > Attach source tables from a connected data source as datasets, making them available for analysis and queries. Workflow for browsable sources (warehouses, databases, files): browse the connection's resources first, then pass each chosen ``from_source`` value here exactly as returned — never rewrite or guess source paths. For request-shaped sources (HTTP API connections), omit ``from_source`` and describe the request via ``params`` (``http_request_path``, ``request_body``, ``pagination_data_pointer``, pagination settings …); the server derives the source from the connection. Up to 100 datasets attach in one atomic request; dataset names are derived automatically (with collision handling) unless provided. Attachment returns immediately; data loading continues in the background. Each returned dataset carries a ``status`` and a ``proxy_table_status``; a dataset is queryable only once ``status`` is ``DEPLOYED`` **and** ``proxy_table_status`` is ``CREATED``. Poll the connection's dataset list until both hold before querying. Credentials come from the connection itself — never include secrets in this request. ## OpenAPI ````yaml https://api.summation.com/openapi.json post /v1/connections/data/{connection_id}/datasets openapi: 3.1.0 info: title: summation-api public API version: 0.1.0 servers: [] security: [] tags: - name: Auth description: Machine authentication for API clients. - name: Tenant description: Organization and tenant context resolved by summation-api. - name: Projects description: Project resources. - name: Conversations description: Conversation and message resources. - name: Reports description: Report resources and report operations. - name: Runs description: Project run history and execution status. - name: Playbooks description: Playbook resources. - name: Files description: Project file resources. - name: Catalog Entries description: Project-scoped data asset attachments. - name: Data Connectors description: Reusable external data source connections. - name: App Connectors description: External app connectors whose tools the agent can use during chat. - name: Tables description: Canonical table metadata and catalog. - name: Views description: Canonical Summation view metadata and catalog. - name: Query Executions description: Read-only SQL query execution. - name: Grid description: Materialized grid status, sync, and lineage. - name: Schedules description: Schedule and schedule run inspection. paths: /v1/connections/data/{connection_id}/datasets: post: tags: - Data Connectors summary: Attach connection datasets description: >- Attach source tables from a connected data source as datasets, making them available for analysis and queries. Workflow for browsable sources (warehouses, databases, files): browse the connection's resources first, then pass each chosen ``from_source`` value here exactly as returned — never rewrite or guess source paths. For request-shaped sources (HTTP API connections), omit ``from_source`` and describe the request via ``params`` (``http_request_path``, ``request_body``, ``pagination_data_pointer``, pagination settings …); the server derives the source from the connection. Up to 100 datasets attach in one atomic request; dataset names are derived automatically (with collision handling) unless provided. Attachment returns immediately; data loading continues in the background. Each returned dataset carries a ``status`` and a ``proxy_table_status``; a dataset is queryable only once ``status`` is ``DEPLOYED`` **and** ``proxy_table_status`` is ``CREATED``. Poll the connection's dataset list until both hold before querying. Credentials come from the connection itself — never include secrets in this request. operationId: attach_data_connection_datasets parameters: - name: connection_id in: path required: true schema: type: string title: Connection Id requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/ConnectionDatasetsAttachRequest' responses: '200': description: Successful Response content: application/json: schema: {} '400': description: Invalid request or unsupported option. content: application/problem+json: schema: $ref: '#/components/schemas/ProblemError' '401': description: Missing, invalid, or expired credentials. content: application/problem+json: schema: $ref: '#/components/schemas/ProblemError' '403': description: Authenticated but missing the required scope. content: application/problem+json: schema: $ref: '#/components/schemas/ProblemError' '404': description: Resource not found or not accessible. content: application/problem+json: schema: $ref: '#/components/schemas/ProblemError' '422': description: Validation Error content: application/json: schema: $ref: '#/components/schemas/HTTPValidationError' '429': description: Rate limit exceeded. content: application/problem+json: schema: $ref: '#/components/schemas/ProblemError' '500': description: Internal server error. content: application/problem+json: schema: $ref: '#/components/schemas/ProblemError' security: - SummationMachineAuth: - agent:write components: schemas: ConnectionDatasetsAttachRequest: properties: datasets: items: $ref: '#/components/schemas/DatasetAttachSpec' type: array maxItems: 100 minItems: 1 title: Datasets description: >- Datasets to attach — 1 to 100 per request, applied as one atomic batch. type: object required: - datasets title: ConnectionDatasetsAttachRequest ProblemError: description: >- Public error envelope (RFC 7807 ``application/problem+json``). Mirrors the body produced by :func:`api_error`. Declaring it in the contract lets generated clients type error handling instead of guessing. properties: type: description: URI reference identifying the problem type. title: Type type: string title: description: Short, human-readable summary of the problem type. title: Title type: string status: description: HTTP status code. title: Status type: integer detail: description: Human-readable explanation specific to this occurrence. title: Detail type: string code: description: Stable, machine-readable error code. title: Code type: string request_id: description: Correlation id, echoed in the x-request-id header. title: Request Id type: string required: - type - title - status - detail - code - request_id title: ProblemError type: object HTTPValidationError: properties: detail: items: $ref: '#/components/schemas/ValidationError' type: array title: Detail type: object title: HTTPValidationError DatasetAttachSpec: properties: from_source: anyOf: - type: string minLength: 1 - type: 'null' title: From Source description: >- Source path exactly as returned by browse connection resources. Omit for request-shaped connections (e.g. HTTP APIs), where the server derives it from the connection and params. name: anyOf: - type: string - type: 'null' title: Name description: Optional dataset name; the server derives one when omitted. description: anyOf: - type: string - type: 'null' title: Description description: Optional human-readable description for the dataset. snapshot_enabled: anyOf: - type: boolean - type: 'null' title: Snapshot Enabled description: >- Opt this dataset into snapshotting (a periodic copy into Summation's lakehouse) regardless of the connection's policy. Omit to inherit the connection's snapshot setting. Required before a dataset snapshot can be taken for sources whose connection has snapshotting off. params: anyOf: - additionalProperties: anyOf: - type: string - type: integer - type: number - type: boolean - additionalProperties: anyOf: - type: string - type: integer - type: number - type: boolean type: object type: object - type: 'null' title: Params description: >- Connector-specific dataset parameters, using exactly the keys the connection type documents — e.g. for HTTP API connections: http_request_path (required, starts with '/'), request_body (string template), request_headers (flat object), pagination_data_pointer (JSON pointer to the response rows), plus the connector's pagination settings. Values are scalars or one flat object of scalars. Never include credentials — authentication comes from the connection, and auth-related keys are rejected. type: object title: DatasetAttachSpec ValidationError: properties: loc: items: anyOf: - type: string - type: integer type: array title: Location msg: type: string title: Message type: type: string title: Error Type type: object required: - loc - msg - type title: ValidationError securitySchemes: SummationMachineAuth: type: oauth2 flows: clientCredentials: scopes: agent:read: agent:read agent:write: agent:write tokenUrl: /v1/auth/m2m/token ````