Resolving Duplicate Profiles Caused by Uppercase Email Values in AEP | Adobe Higher Education
Skip to main content
Level 1
August 5, 2026
質問

Resolving Duplicate Profiles Caused by Uppercase Email Values in AEP

  • August 5, 2026
  • 1 の返信
  • 31 ビュー

In one of our Web SDK implementations in AEP, we are collecting form data through an HTTP API, and the prospect email ID is being used as the primary identity. During the initial data flow mapping, we did not apply a lowercase transformation to the email field, so email values entered in uppercase were ingested into CDP as-is.

We are now implementing lowercase mapping for the email value in the data flow. However, this is creating an issue because the historical records already stored in uppercase are resulting in separate profiles in CDP. To avoid duplicate profiles, we would like to apply lowercase mapping to the historical data as well in the AEP Profile UI.

Could you please guide us the best approach on how to do this? other than reingestion 

    1 の返信

    Level 1
    August 5, 2026

    Hi Robert, 

    You can use AEP query service to identify all the emails for which you want to fix this and create a single AEP profile. Something like this SELECT DISTINCT email FROM <dataset> WHERE email <> LOWER(email)

    Then you can build a new schema with just two attributes, the uppercase email and lowercase email. And make sure to mark both as identity and ‘Email’ as the namespace.

    I know you have asked for solution without reingestion but this much you would have to do.


    Batch-ingest that small lookup dataset (one row per affected person, not the full event history). Because both values sit on the same record under the same namespace, Identity Service will link them into one identity graph.

    Once ingested, your merge policy pulls profile fragments and events from every identity in that graph, so segmentation, activation, and the Profile UI all show one unified person. The underlying historical event records never get touched or reprocessed.

    Thanks!