Real-Time Profile Lookup

Profile lookup helps you investigate how Real-Time ID stitching resolved a profile. A Treasure Workflow submits one or more identity selectors, and Real-Time writes a result for each input row to the profile_lookup table in the parent segment's real-time database, cdp_audience_<parent_segment_id>_rt.

Use lookup for targeted debugging. It is not intended for exporting the ID graph in bulk. The result contains the IDs currently associated with the resolved profile and its persisted real-time attributes. It does not include batch attributes or dynamically evaluated attributes.

If a profile was merged into another profile, the IDs moved to the retained RID have timestamps corresponding to the merge time. These timestamps describe when an association was updated; they are not necessarily when an ID was first seen.

When multiple selectors in one input row resolve to different profiles, the lookup returns the oldest RID, which is the RID a stitching operation would retain. Use the fewest selectors needed to identify the profile.

You need the parent segment ID, its Real-Time instance ID, and the importer endpoint for your deployment stage and site. The importer endpoint is reachable from Treasure Workflow within the Treasure AI platform. Contact Treasure AI Support if you need the instance ID or help identifying the correct endpoint.

Store an API key in the workflow project's Secrets tab as td.apikey. For this workflow, the secret value must be a complete authorization value in the form TD1 <account_id>/<api_key>.

The following example looks up one profile by its configured td_client_id key. It uses the production US endpoint as an example. Replace XXXX, the instance ID, and the endpoint to match your deployment stage and site. Consult Treasure AI Support if you are unsure which endpoint to use. The optional td_rt_lookup_id value makes it easier to find the result later.

_export:
  reactor_api_endpoint: "https://bulk-storage-importer-api-production-aws-us-east-1.internal.treasuredata.com"
  reactor_instance: "aYYYYnZ"
  td:
    database: cdp_audience_XXXX
    engine: presto

+profile_lookup:
  td>:
  query: |
    SELECT
      'td_client_id:12345' AS result,
      'td_rt_lookup_id:my-lookup-job' AS lookup_id
  result_url: |
    {
      "type": "rest",
      "method": "POST",
      "authorization": "${secret:td.apikey}",
      "endpoint": "${reactor_api_endpoint}/internal/lookup-profile",
      "headers": "{\"x-reactor-instance-name\": \"${reactor_instance}\"}"
    }

Use a key configured for your Real-Time instance. Each selector has the form key:value; the optional td_rt_lookup_id:value is metadata, not a selector. The request accepts up to five input rows. Run targeted lookups at approximately one request per second per importer instance rather than using this workflow for bulk extraction.

After the workflow completes, query the results in Data Workbench:

SELECT
  time,
  td_rt_lookup_id,
  reactor_instance_id,
  lookup_row_index,
  lookup_selectors,
  current_rid,
  realtime_attributes,
  current_id_attributes,
  current_id_timestamps
FROM cdp_audience_XXXX_rt.profile_lookup
WHERE td_rt_lookup_id = 'my-lookup-job'
ORDER BY time DESC

These endpoints apply to production. Other deployment stages use different endpoints; use the endpoint that matches both your stage and site.

For the identity changes that led to the current graph, see Real Time Identity Log Tables. Profile lookup shows the current state; the identity log tables show changes over time.