Check your eligibility for this 50% exam voucher offer and join us for free live learning sessions to get prepared for Exam DP-700.
Get StartedDon't miss out! 2025 Microsoft Fabric Community Conference, March 31 - April 2, Las Vegas, Nevada. Use code MSCUST for a $150 discount. Prices go up February 11th. Register now.
Hi Folks,
I got stuck with some strange fallback issue.
I created Lakehouse with small table with two columns (City, ID). This table have only two rows
I created a new semantic model using SQL endpoint including table which we have in default dataset, i didn't make not even one changes at all.everything is same.
when I create a reports using these semantic models the default semantic model is acting as DirectLake mode, where as newly created semantic model acting as Direct Query Mode.
Using SQl profiler I gathered some traces. Based on traces I have observed that New Semantic model is hitting Lakehouse table as direct query. Please see the screenshot below for more information related to this issue.
New Semantic Model Traces:
When I compare the datasets with ALM toolkit what I have observed is its changing dataformat and aggregation in new semantic model. Also partition names seems to be different.
I tried modifying the Datatypes and Aggs, but as per my understnding it further leads to fallback.
Am I really missing any important point or option? Please suggest.
Thank you.
Regards
BVK
Solved! Go to Solution.
Ohooo I think I found the issue 🕺🕺.
Looks like I missed one last suggestion by MS which might leads to Fallback if we don't follow.
https://learn.microsoft.com/en-us/power-bi/enterprise/directlake-overview
When you create New Semantic Model, even though following option (highlighted) enabled on New Semantic Model still we need to refresh manually as highlighted. As soon as I have done that all queries turned into Dirctlakemode and no Fallback is happening.
Before Refresh :
After Refresh :
Please correct me if I am not understanding it properly. Hope this helps someone as well.
Thanks again for your inputs.
Hi @itsmebvk ,
I do not have a quick solution at hand, but I spontanteously wondered wgether you are querying a table directly or a view (views do not work with Direct Lake)? Another reason could be security (RLS or OLS) that triggers the Direct Query fallback.
Also there is a DMV called TMSCHEMA_DELTA_TABLE_METADATA_STORAGES that has an attribute called FallbackReason that might give you some clue on what's going on 🙂
Hope this helps (at least a bit) 🙂
/Tom
https://www.tackytech.blog/
https://www.instagram.com/tackytechtom/
Did I answer your question❓➡️ Please, mark my post as a solution ✔️ |
Also happily accepting Kudos 🙂 |
Feel free to connect with me on LinkedIn! |
#proudtobeasuperuser |
Hi @tackytechtom ,
Thanks for your quick reply.
To answer your observations.
1) I am not using views, I am using table
2) I didn't apply security or RLS/OLS either in Lakhouse or Semantic Modle layer.
3) I have already tried DMV you have suggested but its not showing details of object which is leading to fallback however its just showing 4 as fallback reasons for new semantic model.
select * from $SYSTEM.TMSCHEMA_DELTA_TABLE_METADATA_STORAGES
Default Semantic Model (Showing 0 Fallback reasons)
New Semantic Model (Showing 4 Fallback reasons) but no additional details available
May be I am wrong, I am assuming its highlighting following four difference which happened automatically.
As far as I know I strictly followed MS suggestions to avoid any fallback but still its turning into Directquery. 😞 😞 .
Any other inputs ?
Ohooo I think I found the issue 🕺🕺.
Looks like I missed one last suggestion by MS which might leads to Fallback if we don't follow.
https://learn.microsoft.com/en-us/power-bi/enterprise/directlake-overview
When you create New Semantic Model, even though following option (highlighted) enabled on New Semantic Model still we need to refresh manually as highlighted. As soon as I have done that all queries turned into Dirctlakemode and no Fallback is happening.
Before Refresh :
After Refresh :
Please correct me if I am not understanding it properly. Hope this helps someone as well.
Thanks again for your inputs.
March 31 - April 2, 2025, in Las Vegas, Nevada. Use code MSCUST for a $150 discount!
Check out the January 2025 Power BI update to learn about new features in Reporting, Modeling, and Data Connectivity.
User | Count |
---|---|
25 | |
17 | |
10 | |
10 | |
9 |
User | Count |
---|---|
35 | |
26 | |
18 | |
16 | |
13 |