Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
Important
This feature is in Beta. To use it, a workspace admin must turn on Lakeflow Connect for Vimeo from the Previews page. See Manage Azure Databricks previews.
The managed Vimeo connector in Lakeflow Connect supports four source tables in the default schema. The teams, team_members, and team_roles tables are full refresh only, and audit_logs syncs incrementally on created_time. The following sections list each table's primary key, the start_datetime option, and the destination table schemas.
Supported source tables
The Vimeo connector supports the following source tables, all under the default source schema:
| Source table | Primary key | Description | Sync mode | Cursor field |
|---|---|---|---|---|
teams |
id |
Teams that the account owner belongs to, including team size and the token user's role on the team. | Full refresh only | |
team_members |
uri |
Members of the account owner's team, including role, permission level, and membership status. | Full refresh only | |
team_roles |
uri |
Permission policies (team roles) defined for the account owner's team. | Full refresh only | |
audit_logs |
uri |
Audit log events for the Enterprise team account, including the actor, event, and target object. | Incremental | created_time |
Connector options
The Vimeo connector supports the following pipeline-level configuration option. See Examples for usage.
| Option | Type | Scope | Required | Applies to | Description |
|---|---|---|---|---|---|
start_datetime |
string | Pipeline | No | audit_logs |
The earliest date (yyyy-MM-dd) to ingest audit log events from on the first sync. Defaults to 364 days before the first sync. Earlier dates are clamped to 364 days, just under Vimeo's 365-day limit on audit log ranges. Later syncs continue from the last ingested event. |
Destination table schemas
The following sections describe the schema of each destination table. None of the tables support SCD type 2 because each table has VARIANT columns.
teams
Primary key: id
Sync mode: Full refresh only
| Field | Data type |
|---|---|
id |
BIGINT |
team_data |
VARIANT |
current_team_size |
BIGINT |
max_team_size |
BIGINT |
owner |
VARIANT |
user_role |
STRING |
untranslated_user_role |
STRING |
has_team_management |
BOOLEAN |
has_content_shared |
BOOLEAN |
team_members
Primary key: uri
Sync mode: Full refresh only
| Field | Data type |
|---|---|
uri |
STRING |
role |
STRING |
permission_level |
STRING |
invite_url |
STRING |
email |
STRING |
user |
VARIANT |
created_time |
TIMESTAMP |
modified_time |
TIMESTAMP |
joined_time |
TIMESTAMP |
status |
STRING |
active |
BOOLEAN |
resource_key |
STRING |
metadata |
VARIANT |
access_grant |
STRING |
team_roles
Primary key: uri
Sync mode: Full refresh only
| Field | Data type |
|---|---|
uri |
STRING |
name |
STRING |
created_on |
TIMESTAMP |
modified_on |
TIMESTAMP |
permission_actions |
VARIANT |
audit_logs
Primary key: uri
Cursor field: created_time
| Field | Data type |
|---|---|
uri |
STRING |
event_name |
STRING |
event_name_translated |
STRING |
created_time |
TIMESTAMP |
actor_user_name |
STRING |
actor_user_email |
STRING |
actor_user_role |
STRING |
actor_user_pictures |
VARIANT |
actor_location |
STRING |
actor_platform |
STRING |
actor_ip |
STRING |
by_staff |
BOOLEAN |
object |
STRING |
object_id |
BIGINT |
object_link |
STRING |
privacy_view |
STRING |
title |
STRING |
workspace_display_name |
STRING |
details |
VARIANT |
Required Vimeo account permissions
The connector requires the following Vimeo account permissions and settings.
| Permission | Required for |
|---|---|
| Account Owner or Administrator of the team | Generate the personal access token |
Personal access token with the public, private, and stats scopes |
Baseline access to teams, team members, and team roles |
| Enterprise Vimeo account with audit logs enabled, and an API app created on that account | audit_logs |