Incremental Refresh question -help

[srk_powerbi]

Hi, 

i have a question about how incremental refresh works.

1. can we implement incremental refresh without a Date field in fact table? and My table does not have date field and, only has datekey in the format of integer number its not in the format 'mmddyyyy'. 

2. There are last 10 yrs data so Initially i load  data from 2018 and above and set up incremental refresh. But in future, i i want data from before 2018, can we load that?

please suggest

[pleblanc]

Gotcha, so what does 1 equal to from a time perspective, what does 2 equate to?  etc..

[pleblanc]

Yes you can, watch this video https://youtu.be/wjdMARaedcU.

[srk_powerbi]

@pleblanc  yes for both my questions?

[pleblanc]

Yes.  You actually create function to convert the Integers in to dates.

[srk_powerbi]

how do we convert integer to date?   my date key is not in the format mmddyyyy or any other date..its a just a number starting from 1 to  10000

[v-jingzhang · Accepted solution]

Hi @srk_powerbi 

Can you map your dates to actual calendar dates? If yes, you can add a custom column into the fact table which contains calendar dates mapping to original integer dates, then filter the rows based on this new calendar date column. If these integer dates cannot be mapped to calendar dates, I'm afraid it may be impossible to configure incremental refresh. When you publish the report into the service, Power BI will identify the datetime range partition based on the service time. It is not able to understand non-standard dates.

For your second question, when you load data in Power BI Desktop, you can set up any dates for RangeStart and RangeEnd parameters. You can import small data which is enough to create the report. But when you configure incremental refresh, you need to set up store rows in the last 10 years if you want to hold 10 years' data in the dataset. After you publish the report into the service, the first refresh in the Power BI service will import all 10 full calendar years historical data. Subsequent refreshes are incremental and will not refresh historical data. If you don't import enough historical data, you will not be able to load them into the dataset in the future. In that situation, you can only modify the incremental refresh policy in Power BI Desktop and publish the report into the service again to perform a new first refresh.

Reference:

Incremental refresh in Power BI - Power BI | Microsoft Docs

How to configure Incremental Refreshing in Power BI with DateKey or Date (instead of the default Dat...

Kindly let me know if this helps.
Community Support Team _ Jing
If this post helps, please Accept it as the solution to help other members find it.

[srk_powerbi]

@v-jingzhang  the integer keys are like 1,2,3,4 .......etc which are mapped to date filed in date dimension...In this case how do i convert these integers to date format? Does the integers need to be in the format 20210429  or any integer s fine?