DocumentationExperience PlatformSource Connectors Guide

[Beta]{class="badge informative"}

Delta Sharing

Last update: June 16, 2026

CREATED FOR:

  • Developer
AVAILABILITY
This feature is currently in a limited beta and will only be available until July 15, 2026. Contact your Adobe account team to request access to the beta.

With the Delta Sharing source connector, you can securely connect to your Databricks Delta Sharing source and create virtual datasets in Adobe Experience Platform. This allows you to query and use external tables through Experience Platform services, while representing the shared data through relational schemas, without physically ingesting the data into Platform. By using Delta Sharing, you can reduce duplicate data storage, lower storage costs, and simplify data management.

Prerequisites prerequisites

Before you connect to a Databricks Delta Sharing source from Adobe Experience Platform, ensure the following requirements are met.

IMPORTANT
Delta Sharing does not physically ingest or copy source data into Experience Platform. Instead, shared data is represented in Experience Platform as virtual datasets. Ensure that the shared tables you want to use are available through Delta Sharing and can be represented through relational schemas in Experience Platform.

Databricks provider prerequisites

On the provider side, ensure that you have the following:

  • External Delta Sharing is enabled: External Delta Sharing must be enabled for your Databricks account/workspace, and a default token TTL configured.

  • Eligible Delta tables: The tables you plan to share must:

    • Be Delta tables.
    • Be configured on storage that supports open Delta Sharing (external storage or Databricks‑supported configuration).
    • Not have row filters or column masks applied (Databricks does not allow sharing such tables).

Share and recipient configuration

A Databricks admin must complete the following steps:

  1. Create a share and add tables:

    code language-sql 
    CREATE SHARE IF NOT EXISTS my_first_share
  COMMENT 'Share for Adobe Experience Platform';

ALTER SHARE my_first_share
  ADD TABLE {CATALOG}.{SCHEMA}.{TABLE};

  2. Create a recipient for Experience Platform and grant the share:

    code language-sql 
    CREATE RECIPIENT IF NOT EXISTS plat_recipient
  COMMENT 'Recipient for Platform Delta Sharing source';
GRANT SELECT ON SHARE my_first_share TO RECIPIENT plat_recipient;

Download the Delta Sharing credentials file (.share) for this recipient. This file contains the four values you will use in the Experience Platform UI.

Select to view an example .share file
code language-json 
{
  "shareCredentialsVersion": 1,
  "endpoint": "https://{WORKSPACE}.azuredatabricks.net/api/2.0/delta-sharing/metastores/{ID}",
  "bearerToken": "dapi1234567890abcdef1234567890abcdef-1234567890abcdef",
  "expirationTime": "2026-03-31T23:59:59Z"
}

Network and firewall configuration

If your cloud storage accounts such as, Azure Data Lake Storage Gen2, Amazon S3, Google Cloud Storage are protected by firewall rules, ensure that Experience Platform can access the required Delta Sharing resource, including:

  • The Delta Sharing endpoint from the .share file.
  • The underlying storage paths for the shared tables.

Experience Platform prerequisites

Supported environment

  • Your Experience Platform organization must be hosted in a supported Azure region.
  • The Delta Sharing source feature must be enabled for your organization. If you do not see the “Data Sharing → Databricks Delta Share” card in the Sources catalog, contact Adobe Support or your Adobe representative.

Additionally, the user creating the connection must have the following permissions in the target sandbox:

  • Access to Sources and Permission to create source connections / dataflows.
  • Permission to create schemas and datasets, including virtual datasets.

Governance expectations

  • Delta Sharing connections create virtual datasets in Catalog. These datasets are read-only and are represented as virtual datasets in Experience Platform.
  • No rows are ingested into Experience Platform. Only metadata, such as schema, lineage, and connection details, is stored.
  • Standard governance labels and policies can be applied to the virtual schema, but privacy/retention jobs do not modify the source data.

Gather required credentials gather-required-credentials

You must provide values for the following credentials to authenticate and use the Delta Sharing source:

Credential
Description
Example
Endpoint
The Endpoint is the base URL of the Delta Sharing server that hosts the shared tables. Generated by Databricks when external Delta Sharing is enabled, this endpoint appears in the recipient’s .share credentials file and is used by Experience Platform to enumerate shares, browse schemas and tables, and fetch metadata and data previews.
https://adb-1234567890123.4.azuredatabricks.net/api/2.0/delta-sharing/metastores/0a1b2c3d-4e5f-6789-abcd-0123456789ef
Bearer token
The Bearer token is a read-only Delta Sharing access token associated with a specific recipient in Databricks. It authenticates Experience Platform as the recipient to the Delta Sharing server by being included as an Authorization header in every request (Authorization: Bearer {BEARER_TOKEN}).
dapi1a2b3c4d5e6f7g8h9i0jklmnopqrstuvwx
Share credentials version
Share credentials version corresponds to shareCredentialsVersion in the .share file. It is the version of the credentials file schema, not a token version. Experience Platform uses it to understand how to interpret the fields in the profile. For today’s Databricks Delta Sharing profiles, the value is 1.
1
Expiration time
Expiration time is an optional timestamp specifying when the bearer token will expire. Expiration time must be in a valid UTC ISO‑8601 timestamp.
2026-03-31T23:59:59Z

Allowed XDM field data types allowed-xdm-field-data-types

When mapping shared tables to Experience Platform schemas, only the following XDM field data types are allowed:

  • Core primitives: String, Number, Integer, Boolean
  • Sub-types via constraints: Byte, Short, Integer (32‑bit), Long
  • Temporal: Date, DateTime
  • Structural: Array, Object
  • Constrained string variants: Enum / Suggested list (meta:enum)

Connect to Databricks Delta Sharing in the UI

Read the Databricks Delta Sharing UI guide to learn how you can ingest data to Experience Platform with the Delta Sharing source.

recommendation-more-help
experience-platform-help-sources