Forum Discussion
Helper IFormula.Firewall Error which disappears after reloading the data model in PBI Desktop
Hi there,
I was wondering if anyone could help me after I did not find any solution that helped me removing the following error:
Formula.Firewall: Query 'Query1' (step 'Source') references other queries or steps, so it may not directly access a data source. Please rebuild this data combination.
I already read the following topics on this topic but none had the solution. Except, I failed to execute the instructions:
1. https://docs.microsoft.com/en-us/power-query/dataprivacyfirewall (Every merging step is executed in a seperate query)
2. https://excelguru.ca/power-query-errors-please-rebuild-this-data-combination/ (All data is loaded into a staging area before merging steps take place, disabling privacy levels are still activated, deactivating did not help)
3. https://www.thebiccountant.com/2020/11/16/improve-file-import-from-sharepoint-in-power-bi-and-power-query/ (I rebuild my SharePoint query to enhance the performance)
I also changed the data type from excel to csv. I am retrieving data solely from two SharePoint 365 folders with several subfolders (privacy level: organisation). I reshaped my query a dozen times and am still recieving the Error after loading my model in PBI Desktop for >15 minutes. After reloading within Power BI Desktop without any changes made the data model will be loaded.
If anyone of you guys has an idea what I am missing, please help me out. I can provide you more information if needed.
Regards
KS
6 Replies
- lbendlin
Super User
(Every merging step is executed in a seperate query)Per my current understanding (and I may be wrong) that is the issue. To avoid the formula firewall you should keep everything in the same partition.
KimStahl94Hi @lbendlin,
thanks for your answer. In my understanding a partition is created after a merging step is executed. That is why I changed my query from the single query (left) to the three queries (right).
How could I keep it all in the same partition if a new partiton is created after the merge? Or am I getting the term partitions wrong?
Important note: Privacy Levels are all set to organisational. Shouldn't the firewall not be a problem then?Regards
KS- lbendlin
Avoid merging at all cost. Either do the joins further upstream, or use separate dataflows, or let the data model do the work for you.
