Forum Discussion

PieterVeenstra's avatar
PieterVeenstra
New Member
4 months ago
Solved

Custom connectors in Fabric Dataflows

I've got the following setup:

Lakehouse that is filled by data from a custom api. 

 

The custom api asks for a username and password and then returns the access token. The access token expires when a new accesstoken is requested. This means that if I connect to the api with Web.Content calls that I run into issues with authentication. 

 

Now I've read up about custom connectors being a solution to this.

 

I managed to make a custom connector work within my Power BI report. But from the Dataflow I can't see a way to make this custom connector available. Apparently through an on-premises data gateway is an options but that feels like not a great option either.

 

What is the best approach for this?

  • I have found a slighlty easier solution. I am importing data into about 10 table from a dodgy API. Some of the tables need frequent updates, other can do with updates on a daily basis. I split my dataflows into 10 dataflows each handling one table. Then through scheduling them cleverly (the daily ones go through the night and the important ones during the day on a more frequent schedule) I managed to get all the data processed.

4 Replies

  • Apparently through an on-premises data gateway is an options but that feels like not a great option either.

    That is your ONLY option.

  • v-priyankata's avatar
    v-priyankata
    Community Support

    Hi PieterVeenstra 

    Thank you for reaching out to the Microsoft Fabric Forum Community.

    lbendlin Thanks for the inputs.

    I hope the information provided by user was helpful. If you still have questions, please don't hesitate to reach out to the community.

     

  • I have found a slighlty easier solution. I am importing data into about 10 table from a dodgy API. Some of the tables need frequent updates, other can do with updates on a daily basis. I split my dataflows into 10 dataflows each handling one table. Then through scheduling them cleverly (the daily ones go through the night and the important ones during the day on a more frequent schedule) I managed to get all the data processed.

    • v-priyankata's avatar
      v-priyankata
      Community Support

      Hi PieterVeenstra 
      Thanks for the letting us know this.
      If your issue has been resolved, it could help others in the community facing the same problem. Please mark the helpful reply or your own reply as the accepted solution.

      Thanks.