Record delete requests (UI Workflow) record-delete

Use the Data Lifecycle workspace to delete records in Adobe Experience Platform based on their primary identities. These records can be tied to individual consumers or any other entity that is included in the identity graph.

Prerequisites prerequisites

Deleting records requires a working understanding of how identity fields function in Experience Platform. Specifically, you must know the identity namespace values of the entities whose records you want to delete, depending on the dataset (or datasets) you are deleting them from.

Refer to the following documentation for more information on identities in Experience Platform:

Create a request create-request

To start the process, select Data Lifecycle in the left navigation of the Experience Platform UI. The Data lifecycle requests workspace appears. Next, select Create request from the main page in the workspace.

The request creation workflow appears. By default, the Delete record option is selected under the Requested Action section. Leave this option selected.

Select datasets select-dataset

The next step is to determine whether you want to delete records from a single dataset or all datasets. Depending on your organization’s configuration, the dataset selection option may not be available. If you do not see this option, continue to the Provide identities section of the guide.

In the Record Details section, select a radio button to choose either a specific dataset or all datasets.

To delete from a specific dataset, select Select dataset, then select the database icon ( ). In the dialog that appears, choose a dataset and select Done to confirm.

To delete from all datasets, select All datasets. This option increases the scope of the operation and requires you to provide all relevant identity types.

Each dataset in Experience Platform supports only one primary identity type.

Provide identities provide-identities

When deleting records, you must provide identity information so the system can determine which records are to be deleted. For any dataset in Experience Platform, records are deleted based on the identity namespace field that is defined by the dataset’s schema.

Like all identity fields in Experience Platform, an identity namespace is composed of two things: a type (sometimes referred to as an identity namespace) and a value. The identity type provides context as to how the field identifies a record (such as an email address). The value represents a record’s specific identity for that type (for example, jdoe@example.com for the email identity type). Common fields used as identities include account information, device IDs, and cookie IDs.

TIP

If you don’t know the identity namespace for a particular dataset, you can find it in the Experience Platform UI. In the Datasets workspace, select the dataset in question from the list. On the details page for the dataset, hover over the name of the dataset’s schema in the right rail. The identity namespace is displayed along with the schema name and description.

There are two options to provide identities when deleting records:

Upload a JSON file upload-json

To upload a JSON file, you can drag and drop the file into the provided area, or select Choose files to browse and select from your local directory.

The JSON file must be formatted as an array of objects, each object representing an identity.

[
  {
    "namespaceCode": "email",
    "value": "jdoe@example.com"
  },
  {
    "namespaceCode": "email",
    "value": "san.gray@example.com"
  }
]

Once the file is uploaded, you can continue to submit the request.

Manually enter identities manual-identity

To enter identities manually, select Add identity.

Controls appear that allow you to enter identities one at a time. Under identity namespace, use the dropdown menu to select the identity type. Under Primary Identity Value, provide the identity namespace value for the record.

To add more identities, select the plus icon ( ) next to one of the rows, or select Add identity.

Quotas and processing timelines quotas

Record Delete requests are subject to daily and monthly identifier submission limits, determined by your organization’s license entitlement. These limits apply to both UI- and API-based delete requests.

Monthly submission entitlement by product quota-limits

The table below outlines identifier submission limits by product and entitlement level. For each product, the monthly cap is the lesser of two values: a fixed identifier ceiling or a percentage-based threshold tied to your licensed data volume.

Quotas reset on the first day of each calendar month. Unused quota does not carry over.

Processing timelines for identifier submissions sla-processing-timelines

After submission, record delete requests are queued and processed based on your entitlement level.

If your organization requires higher limits, contact your Adobe representative for an entitlement review.

Submit the request submit

Once you have finished adding identities to the request, under Request settings, provide a name and optional description for the request before selecting Submit.

A Confirm request dialog appears to indicate that the identities cannot be recovered once deleted. Select Submit to confirm the list of identities whose data you want to delete.

After the request is submitted, a work order is created and appears on the Record tab of the Data Lifecycle workspace. From here, you can monitor the work order’s status as it processes the request.

Deleting records from datasets based on relational schemas relational-record-delete

If the dataset you are deleting from is based on a relational schema, review the following considerations to ensure records are removed correctly and not re-ingested due to mismatches between Experience Platform and your source system.

Record deletion behavior

The following table outlines how record deletions behave across Experience Platform and source systems, depending on the ingestion method and change data capture configuration.

To prevent re-ingestion, apply the same deletion approach in both your source system and Experience Platform, either by removing records from both systems or including _change_request_type = 'd' for records you intend to delete.

Change data capture and control columns

Relational schemas that use Sources with change data capture can use the _change_request_type control column when distinguishing deletes from upserts. During ingestion, records flagged with d are deleted from the dataset, while those flagged with u or without the column are treated as upserts. The _change_request_type column is read at ingestion time only and is not stored in the target schema or mapped to XDM fields.

Additional deletion methods for relational schemas

Beyond the standard record deletion workflow, relational schemas support additional methods for specific use cases:

Descriptor support for hygiene operations descriptor-support

Relational schema descriptors provide essential metadata for precise hygiene operations:

Scheduled retention for relational schemas

For automated hygiene based on data age rather than specific identities, see Manage Experience Event dataset retention (TTL) for scheduled row-level retention in the data lake.

Best practices for relational record deletion

To avoid unintentional re-ingestion and maintain data consistency across systems, follow these best practices:

For more details on schema requirements, see relational schema descriptor requirements.

To learn how change data capture works with sources, see Enable change data capture in sources.

Next steps

This document covered how to delete records in the Experience Platform UI. For information on how to perform other data lifecycle management tasks in the UI, refer to the Data Lifecycle UI overview.

To learn how to delete records using the Data Hygiene API, refer to the work order endpoint guide.