Dual table strange behavior

[MARUI0827]

My model has a dimension table with dual storage mode A and a compute table B,I created A 1-to-many one-way relationship for A and B,like this:

The query is:

EVALUATE
SUMMARIZECOLUMNS(
	'Dim B'[ProdKey],
	"result",[result]
)

I found that the performance was not good enough.

By checking the query plans, I found that SE did not directly aggregate results based on the dimensions of table B,instead, two xmsql's are executed.

The first xmsql queries the prodkey of table B,the second xmsql aggregated results from the dimensions of table A, and then in FE,executed TreatAs and CrossApply:

When I create a table C whose storage mode is also DUAL,FE does not require executed  TreatAs and CrossApply,it becomes very simple.

I don't understand why this is happening,I would like to ask if anyone has encountered a similar situation and knows the reason. Thank you very much for your reply.

[foodd]

Hello @MARUI0827 , and thank you for sharing a question with the Commuity.  The following is informational.  Please remember to adhere to the decorum of the Community Forum when asking a question.

Please provide your work-in-progress Power BI Desktop file (with sensitive information removed) that covers your issue or question completely in a usable format (not as a screenshot). You can upload the PBIX file to a cloud storage service such as OneDrive, Google Drive, Dropbox, or to a Github repository, and then share a file’s URL.

https://community.fabric.microsoft.com/t5/Community-Blog/How-to-provide-sample-data-in-the-Power-BI-...

Please show the expected outcome based on the sample data you provided.

https://community.fabric.microsoft.com/t5/Desktop/How-to-Get-Your-Question-Answered-Quickly/m-p/1447...

This allows members of the Forum to assess the state of the model, report layer, relationships, and any DAX applied.