Incremental Refresh vs Default Refresh

[Anonymous]

Hello,

I was trying to explore the incremental refresh functionality on PowerBi since it could be interesting for a client of mine. To test it I did the following:

• I setup a simple PowerBi report in PowerBi Desktop with an Excel file on my OneDrive as the data source
• I created the 'RangeStart' and 'RangeEnd' parameters and filtered the data on a date/time column in the file
• I defined the policy archiving data 2 months before refresh date and incrementally refreshing data 30 days before refresh date

   

• Then I uploaded the report on PowerBi Service

I instantly did the initial refresh (as said on this page) on the model and then I modified my Excel file with some new rows containing different dates. The second refresh was substantially quicker than the first as it went from 25 minutes to only 8.

This is the refresh history on the semantic model with incremental refresh:

Just for testing purpose I tried to create another PowerBi report copying everything from before but without setting up the incremental refresh to see if it really was better then the normal refresh.

I was astonished when the normal refresh with the same report and the same file only took 1 minute to refresh (image below).

This is the refresh history of the semantic model without incremental refresh:

I tried to add even more new rows to the file and it was every time quicker without the incremental refresh.

So my questions are: am i doing something wrong? Is incremental refresh really useful?

The only thing I have in mind is that Excel may not be the best type of data source for incremental refresh, but I couldn't find anything proving that on multiple forums.

Thanks in advance for anyone who'll try to help me

[Akash_Varuna · Accepted solution]

Hi @Anonymous Incrimental refresh is actually usefull for large data set in your particular case Incremental refresh has overhead for comparing and merging data, which isn't beneficial for small flat files like Excel and Default refresh reloads the entire file faster since Excel files are typically loaded in memory as a whole.
In Conclusion Incremental refresh is useful for large datasets where querying only recent changes reduces refresh time. For small flat files, stick to default refresh.
If this post helped please do give a kudos and accept this as a solution
Thanks In Advance