Forum Discussion
Obsolete data sources in dataflow lineage
Hi JenM ,
Actually there's no such functionality to work it out,but I have a workaround,that is, you can click on the bottom right corner which I marked in red circle,then you will see that all related to the current datasource will be highlighted,and others will be grayed out.
Here is the related reference.
https://docs.microsoft.com/en-us/power-bi/collaborate-share/service-data-lineage
Kelly
- JenM6 years agoAdvocate II
Sorry, but your workaround doesn't address the bug. If I click on the arrows for the dataflow, all 7 upstream data sources get highlighted. Only one of them is valid, the others should not exist, and they provide bad intelligence.
My (horrible) workaround is to export the dataflow as json, delete the dataflow and then reimport the json to a new dataflow. When I do this, the lineage diagram correctly shows just a single datasource for the dataflow. Of course this breaks every single usage of the dataflow, so in a high-usage scenario this would be unacceptible.
- v-kelly-msft6 years agoCommunity Support
Hi JenM ,
If so ,I have a simple way for you:
Connect the dateflow in desktop,then go to data source setting and clear your data source then reenter it:
Finally save as a new report or cover the previous report and publish to service.
I guess this way will help solve your issue.
Best Regards,
KellyDid I answer your question? Mark my post as a solution!- JamesDBartlett32 years agoAdvocate II
The question was how to remove old Data Sources from a Power BI Dataflow, not a Report in Power BI Desktop. Please see the answer posted by EG151 above, which actually resolved the issue: Re: Obsolete data sources in dataflow lineage - Microsoft Fabric Community
- v-kelly-msft6 years agoCommunity Support
Hi JenM ,
Is your issue solved?
Best Regards,
KellyDid I answer your question? Mark my post as a solution!- JenM6 years agoAdvocate II
No, the problem is not resolved. Then only way to remove obsolete data sources is to export and reimport the dataflow (as JSON). But that breaks any downstream usage...