Skip to main content
Solved

How to merge initiator attributes with Gemini Connector response in FME?

  • February 4, 2026
  • 2 replies
  • 45 views
pbonagiri
Contributor
Forum|alt.badge.img+5

I’m using the Gemini Connector (Transformer) in FME and facing an issue with initiator (upstream) attributes.

My requirement is:

  • The attributes coming into the Gemini Connector (initiator attributes)
  • Need to be merged with the Gemini response/output, so that the final feature contains both the original attributes and the Gemini result

    Current behavior:

  • The Gemini Connector output only contains response-related attributes
  • Initiator attributes are not carried through or merged automatically

  • Question:

  • Is there a supported way to merge or carry initiator attributes along with the Gemini response?
  • Is this a limitation of the Gemini Connector transformer?
  • What is the recommended pattern (FeatureMerger, Joiner, correlation ID, etc.) to achieve this?

  • Any guidance, best practices, or examples would be really helpful.

    Thanks in advance!

Best answer by crystalwang

Hi ​@pbonagiri,

Thank you for posting your question on the FME Community! I can definitely help suggest a workflow for merging upstream (initiator) attributes with the AI output attributes.

A common pattern for this is to first create or identify a unique ID attribute on your input features. You then pass that unique ID through the AI call in the prompt and use it to merge the AI response back to the original features.

I’ve included a screenshot of an example workspace illustrating this pattern, and I’ve also attached the .fmw in the zip file on this post. In my example I’m using an OpenAIConnector, but the same approach applies to the Gemini Connector as well.

In this setup, the Merged output port of the FeatureMerger produces a final feature that contains:

  • All original (initiator) attributes, and

  • The attributes extracted from the AI response.

Hope this helps! Please feel free to let me know if you have any follow-up questions!

      2 replies

      crystalwang
      Safer
      Forum|alt.badge.img+19
      • crystalwangSafer
      • Safer
      • Best Answer
      • February 5, 2026

      Hi ​@pbonagiri,

      Thank you for posting your question on the FME Community! I can definitely help suggest a workflow for merging upstream (initiator) attributes with the AI output attributes.

      A common pattern for this is to first create or identify a unique ID attribute on your input features. You then pass that unique ID through the AI call in the prompt and use it to merge the AI response back to the original features.

      I’ve included a screenshot of an example workspace illustrating this pattern, and I’ve also attached the .fmw in the zip file on this post. In my example I’m using an OpenAIConnector, but the same approach applies to the Gemini Connector as well.

      In this setup, the Merged output port of the FeatureMerger produces a final feature that contains:

      • All original (initiator) attributes, and

      • The attributes extracted from the AI response.

      Hope this helps! Please feel free to let me know if you have any follow-up questions!

      pbonagiri
      Contributor
      Forum|alt.badge.img+5
      • pbonagiriContributor
      • Author
      • Contributor
      • February 11, 2026

      Hi Crystal Wang,

      Thanks for the Answer I have tested with above Set up and its working

      Thanks,

      Praveen Kumar

        Community Stats

        32,330
        Posts
        123,146
        Replies
        40,364
        Members
        Terms & ConditionsCookie settingsAccessibility statement
        A product of
        Safe Software respectfully acknowledges that we live, learn and work on the traditional and unceded territories of the Kwantlen, Katzie, and Semiahmoo First Nations.