Manage OneLake shortcuts — fabric_onelake

Lists, inspects, creates or updates, and deletes shortcuts on a Fabric item. Discovered Fabric items can be used directly as OneLake targets. A validated raw target list supports connection-backed shortcut types already configured in Fabric without copying data into R.

Usage

fabric_onelake_shortcuts(
  item,
  workspace = NULL,
  item_type = NULL,
  parent_path = NULL,
  tenant_id = Sys.getenv("FABRICQUERYR_TENANT_ID"),
  client_id = Sys.getenv("FABRICQUERYR_CLIENT_ID", unset =
    "04b07795-8ddb-461a-bbee-02f9e1bf7b46"),
  token = NULL,
  auth_args = list(),
  api_base = .fabric_api_base
)

fabric_onelake_shortcut_get(
  item,
  path,
  name,
  workspace = NULL,
  item_type = NULL,
  tenant_id = Sys.getenv("FABRICQUERYR_TENANT_ID"),
  client_id = Sys.getenv("FABRICQUERYR_CLIENT_ID", unset =
    "04b07795-8ddb-461a-bbee-02f9e1bf7b46"),
  token = NULL,
  auth_args = list(),
  api_base = .fabric_api_base
)

fabric_onelake_shortcut_create(
  item,
  path,
  name,
  target,
  workspace = NULL,
  item_type = NULL,
  target_workspace = NULL,
  target_path = NULL,
  target_item_type = NULL,
  conflict_policy = c("Abort", "GenerateUniqueName", "CreateOrOverwrite",
    "OverwriteOnly"),
  tenant_id = Sys.getenv("FABRICQUERYR_TENANT_ID"),
  client_id = Sys.getenv("FABRICQUERYR_CLIENT_ID", unset =
    "04b07795-8ddb-461a-bbee-02f9e1bf7b46"),
  token = NULL,
  auth_args = list(),
  api_base = .fabric_api_base
)

fabric_onelake_shortcuts_bulk_create(
  item,
  shortcuts,
  workspace = NULL,
  item_type = NULL,
  conflict_policy = c("Abort", "GenerateUniqueName", "CreateOrOverwrite",
    "OverwriteOnly"),
  tenant_id = Sys.getenv("FABRICQUERYR_TENANT_ID"),
  client_id = Sys.getenv("FABRICQUERYR_CLIENT_ID", unset =
    "04b07795-8ddb-461a-bbee-02f9e1bf7b46"),
  token = NULL,
  auth_args = list(),
  api_base = .fabric_api_base
)

fabric_onelake_shortcut_cache_reset(
  workspace,
  tenant_id = Sys.getenv("FABRICQUERYR_TENANT_ID"),
  client_id = Sys.getenv("FABRICQUERYR_CLIENT_ID", unset =
    "04b07795-8ddb-461a-bbee-02f9e1bf7b46"),
  token = NULL,
  auth_args = list(),
  api_base = .fabric_api_base
)

fabric_onelake_shortcut_delete(
  item,
  path,
  name,
  workspace = NULL,
  item_type = NULL,
  confirm = FALSE,
  tenant_id = Sys.getenv("FABRICQUERYR_TENANT_ID"),
  client_id = Sys.getenv("FABRICQUERYR_CLIENT_ID", unset =
    "04b07795-8ddb-461a-bbee-02f9e1bf7b46"),
  token = NULL,
  auth_args = list(),
  api_base = .fabric_api_base
)

Arguments

item: Destination Fabric item name, GUID, or object returned by a discovery function.
workspace: Workspace name, GUID, or discovery object containing item. May be omitted when item contains workspaceId.
item_type: Optional item type used to disambiguate a destination item supplied by name.
parent_path: Optional Files or Tables path from which listing starts. Fabric still returns shortcuts below that path exhaustively.
tenant_id: Microsoft Entra tenant ID. Defaults to FABRICQUERYR_TENANT_ID
client_id: Microsoft Entra application/client ID. Defaults to FABRICQUERYR_CLIENT_ID, then the Azure CLI application ID
token: Optional access token or token-provider function. Leave NULL to let 'fabricQueryR' use its normal sign-in flow
auth_args: Additional sign-in options passed to AzureAuth::get_azure_token()
api_base: Fabric REST API base URL. Leave unchanged unless using a different Fabric cloud or a test service
path: Parent Files or Tables path where the shortcut exists or will be created. Local checks validate path syntax, while Fabric applies item- and workload-specific placement rules. For a table shortcut in a Lakehouse without schemas, use Tables. In a schema-enabled Lakehouse, use a schema path such as Tables/dbo. A schema shortcut instead lives under Tables and targets a folder containing multiple Delta tables.
name: Shortcut name.
target: A discovered Fabric item, its name or GUID, or a raw named shortcut target list. A raw target must contain exactly one documented key such as oneLake, adlsGen2, amazonS3, azureBlobStorage, googleCloudStorage, oneDriveSharePoint, s3Compatible, or dataverse. Connection-backed targets must include the documented required fields, use a Fabric connection ID, and must not embed credentials. Local checks cover structure, required fields, identifier shape, and generic URL safety; they do not verify source-specific host/path rules or that a connection refers to the supplied location. A raw oneLake target may also include its documented optional connectionId. Fabric service validation is authoritative.
target_workspace: Workspace containing a OneLake target. May be omitted when a discovered target contains workspaceId.
target_path: Item-relative Files or Tables path for a OneLake target. Required when target is an item and unused for a raw target list.
target_item_type: Optional Fabric item type used to disambiguate a OneLake target supplied by name.
conflict_policy: "Abort" preserves an existing shortcut with the same path and name. "GenerateUniqueName" creates a uniquely named shortcut, "CreateOrOverwrite" creates or updates it, and "OverwriteOnly" updates an existing shortcut without creating one.
shortcuts: For bulk creation, a non-empty list of shortcut request lists. Each request requires path, name, and target, accepts the same target companion fields as fabric_onelake_shortcut_create(), and may include a transform list. The only currently documented transform is csvToDelta; see Details.
confirm: Logical. Deletion is disabled unless explicitly set to TRUE. Deleting a shortcut does not delete its destination data.

Value

fabric_onelake_shortcuts() returns a tibble with one row per shortcut. fabric_onelake_shortcut_get() and fabric_onelake_shortcut_create() return the same one-row shape. fabric_onelake_shortcut_delete() returns TRUE invisibly after success. fabric_onelake_shortcut_cache_reset() returns a fabric_operation handle for either immediate or asynchronous completion.

Details

Shortcut names, parent paths, and OneLake target paths follow Fabric's current shortcut limits: %, +, and non-ASCII characters are rejected. Other source- and destination-specific restrictions are intentionally left to Fabric so that newly supported connection types and rules remain usable.

For example, a table shortcut named orders uses path = "Tables" in a Lakehouse without schemas, or path = "Tables/dbo" in a Lakehouse with schemas; its target_path identifies one Delta table. A schema shortcut named sales uses path = "Tables" and a target such as Tables/sales containing multiple Delta tables. File shortcuts use Files or a folder beneath it and do not register tables.

Listing follows Fabric continuation links and tokens until every shortcut below parent_path is returned. Unknown target details and transform fields are preserved in list columns for forward compatibility.

Create is deliberately not replayed automatically because its POST outcome can be ambiguous after a transport failure. The default conflict policy is Fabric's non-destructive Abort; overwrite must be requested explicitly. Deletion is also not replayed automatically, and a 404 confirms that the shortcut link is already absent.

Bulk creation is a preview Fabric API. Its optional csvToDelta transform accepts includeSubfolders and a properties list containing delimiter, skipFilesWithErrors, and useFirstRowAsHeader. Supported delimiters are comma, space, tab, |, &, and ;. A bulk request returns a fabric_operation; use fabric_operation_result() to retrieve the per-request statuses, created shortcuts, and errors after completion.

These Core REST APIs require OneLake.Read.All or OneLake.ReadWrite.All for reads, and OneLake.ReadWrite.All for create and delete. Fabric documents support for users, service principals, and managed identities. API scope is not sufficient by itself: listing or reading also requires item Read permission or OneLake Read permission on the destination path. Creating requires item Write or OneLake ReadWrite on the destination, plus Read access to the target path. Updating or deleting likewise requires item Write or destination-path OneLake ReadWrite permission.

References

OneLake shortcuts REST API

OneLake shortcut placement and limitations

Create table and schema shortcuts

OneLake shortcut security and path permissions

Create shortcuts in bulk

Reset shortcut cache

Examples

if (FALSE) { # \dontrun{
# Discover two Lakehouses in the same workspace
workspace <- fabric_workspaces()[[1L]]
lakehouses <- fabric_lakehouses(workspace)
destination <- lakehouses[[1L]]
source <- lakehouses[[2L]]
source_paths <- fabric_onelake_list(workspace, source, path = "Tables")
source_table <- source_paths[source_paths$is_directory, ][1L, ]

# Create a shortcut whose target came from the source Lakehouse listing
created <- fabric_onelake_shortcut_create(
  destination,
  path = "Files",
  name = "shared-orders",
  target = source,
  target_path = source_table$path[[1L]]
)

# List the folder, then fetch the created shortcut by its returned identity
fabric_onelake_shortcuts(destination, parent_path = "Files")
shortcut <- fabric_onelake_shortcut_get(
  destination,
  path = created$path[[1L]],
  name = created$name[[1L]]
)

# Delete that same discovered shortcut explicitly
fabric_onelake_shortcut_delete(
  destination,
  path = created$path[[1L]],
  name = created$name[[1L]],
  confirm = TRUE
)
} # }