How semantic model size is increasing?

[jaryszek]

Hello,

I have found in web: 

Explanation:

The semantic model size is measured as the size that is loaded into memory.

When you interact with a direct lake report, the semantic model columns you interact with get loaded into memory (warm up).

So the semantic model will grow in size as report users query more columns from the semantic model.

After a period of inactivity, the columns get removed from memory automatically (cool down).

When you refresh the dataset, alle the columns get removed from memory. So the model size (in memory) will be almost 0 after a refresh.

Can anybody confirm that this is working like that? 
What about oneLake remonte semantic model ?

Best,
Jacek

[Cookistador]

If you use direct lake: yes, it works like that

If you use import or direct query; the size will stay the same (for the import, by default, it should increase after each report)

[jaryszek]

thanks, 

it is any documentation for that? 

"After a period of inactivity," ...what period, how to check this? 

Best,
Jacek

[Cookistador · Accepted solution]

You can find a lot of information in the direct lake documentation:

https://learn.microsoft.com/en-us/fabric/fundamentals/direct-lake-overview

Especially the max size before you roll back to direct query

And here some explanaitions about how it works:

https://learn.microsoft.com/en-us/fabric/fundamentals/direct-lake-understand-storage