This time we’re going bigger than ever. Fabric, Power BI, SQL, AI and more. We're covering it all. You won't want to miss it.
Learn moreDid you hear? There's a new SQL AI Developer certification (DP-800). Start preparing now and be one of the first to get certified. Register now
Hi all,
I have a doubt about the next question. Which is better in terms of performance, Custom Column using PQuery or New Column using DAX?
Thanks to all for you answer.
Solved! Go to Solution.
Yeah, I would say that in your case you should definitely consider moving that to Power Query. General rule is that the further upstream that you can push the calculations, the better the performance downstream.
Well, overall, Power Query is only going to execute the calculation once per a query refreshing, so you will take a small performance hit during query refresh but after that you have a static value. With DAX, the calculation will essentially be done every time it is needed so you will have a performance hit each time.
Ok, thanks so much for your fast answer.
I have a model and i need to split some string values from various columns to use later like slicers. The fact table in the model have more than a million of rows and i'm thinking in transform my new DAX columns in Power Query columns to reduce the performance impact when the user use it.
Do you think i should do it?
Thanks again.
Yeah, I would say that in your case you should definitely consider moving that to Power Query. General rule is that the further upstream that you can push the calculations, the better the performance downstream.
Yes, I thinked the same.
If somebody need it, here you have a link that explain so well my first question:
http://radacad.com/m-or-dax-that-is-the-question
Thanks Greg_Deckler
Check out the April 2026 Power BI update to learn about new features.
Sign up to receive a private message when registration opens and key events begin.
If you have recently started exploring Fabric, we'd love to hear how it's going. Your feedback can help with product improvements.
| User | Count |
|---|---|
| 36 | |
| 33 | |
| 31 | |
| 21 | |
| 16 |
| User | Count |
|---|---|
| 66 | |
| 55 | |
| 31 | |
| 24 | |
| 23 |