Very slow DAX query performance with latest PBI Desktop update

[QC]

I noticed that lately some DAX queries started to run EXTREMELY SLOW -- when I hit return on DAX editor, the system shows "Working on it" message for 20s ~ 3 minutes! Not even fully close-re-open, or restart laptop solve the issue. My dataset wasn't even that large (~200k records). I run into this issue at least twice a day.

I have been using PBI for years, never had this issue before. I think there is something seriously wrong with the latest PBI version. You can see that CPU and Power usage of SSAS are very high during the query.

I am using Version: 2.123.742.0 64-bit (November 2023)

---- Edit----

I mostly encounter this issue when using "ALL" dax function. It seems the engine has issue scanning, which has never happened before.

[v-yifanw-msft]

Hi @QC ,

Based on the information you provided,If you connect in import mode, the query may need a long time to load data into power bi.

 I would recommend you recognize how to optimize queries with this article:

Optimization guide for Power BI - Power BI | Microsoft Learn

If it does not help, please check whether there existing any step in Power Query consuming too much CPU or memory.

1.Try to use Table.Buffer to improve the data load performance.

2.You can also use DAX studio to check out which is the most time-consuming.

Follow this link to install and check:

Use Performance Analyzer to examine report element performance in Power BI Desktop - Power BI | Micr...

Another similar thread could be your reference:

Solved: PBI desktop very slow + freeze - Microsoft Fabric Community

^{Best Regards,}

^{Yifan Wang}

^{If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.}

[QC]

Unfortunately it doesn't solve the issue, I am familiar with all the tools you mentioned, but it's not related to query setup or laptop, all of which remained the same, PBI worked just fine in the past. The issue is related to recent PBI Desktop release. It should not performe so poorly on such small dataset at all. Something is wrong with your engine.