Slowly Changing Dimensions in Power BI

In Power BI we frequently "slice and dice" data using dimensions. For instance, the date dimension allows us to view data by "month", "quarter" or "year". However, dimensions are often fluid; they change over time. This includes products which are slightly refined or employees who move in and out of our business and across the organisational spectrum.
During this session we will use Power BI to explore how to handle these slowly changing dimensions; a key element of the Kimball methodology.
Resources (worked examples, datasets etc) are at bit.ly/33zWQtp in the 'Slowly Changing Dimensions' folder.

[@IngeborgHawighorst]

Thank you for walking through the process in so much detail, and for showing how to do this in Power BI. I have watched many other approaches that go into SQL queries for the changes to the fact table, which I couldn't follow. This was exactly what I needed.

[@shafa7668]

This is one of the most useful vedios I have ever watched on power bi on RUclips.

[@oscarestorach6162]

You are great, thanks a lot to share your knowledge

[@OA-lx2zr]

Best SCD video ive seen

[@shafa7668]

Fantastic vedio. Thanks for doing it.

[@fadwa2413]

Keep Going to learn us ,Please
Perfect as Always

[@pietjehans206]

Thanks, I needed an example of SCD in PBI. Disadvantage of this method is that in order to avoid circular dependency, you need a reference table for each dimension table in your data model. The better option is perhapse to create the index column on the fact table in PowerBI. Perhaps there is an even better solution?

[@ethugp3pp3r]

That is where I think I would have taken a different direction, too. I don’t really like creating Dax tables unless they´re dates…

[@cornel4931]

Hey, thanks so much for this content. The guide clearly knows his way around.
Has any encountered an error with circular dependency?
Since we create the index col on the fact is using the dim table that causes an error when trying to create a relationship.

[@IlariaLongoDorni]

thanks, it solved me a big doubt.

[@LondonBusinessAnalyticsGroup]

Happy to help

[@gauravsingh-jh7pz]

Hi, I was unable to get the total value of JEN (Current revenue ) is there any specific setting we have to do while activating the relationship???

[@TheGlassGalaxy]

thanks sir, is it possible for you to share the link for the tables that you have used so that we can practice the same?

[@LondonBusinessAnalyticsGroup]

here it is github.com/MarkWilcock/lbag-online/tree/master/Slowly%20Changing%20Dimensions

[@dimitridelizee4032]

Thank you for the video. The link seems broken. Can you reload it, please ?

[@pieterlinden8292]

Are the data files available somewhere? The link for them is broken. =(

[@IlariaLongoDorni]

Hi Chris, how could I manage this: I have more than one event (for example 3) in the historical table related to Territory and with the same date. the reference table will correctly show 3 different lines but into the sales table the index number will be only the first one that it founds. I cannot create a bridge table with unique value for historical data because I will lose details that I need. I have tried with rank but many to many relationship are not allowed.thanks

[@chrisbarber639]

This is a hard question to answer without seeing the data. In the case of SCD you want a one to many relationship flowing from the dimension to the fact table. Look at an example record which is causing the many to many, decide on which of those rows which you want to filter the fact table and then look at what else you have in the dimension to distinguish that row.

[@IlariaLongoDorni]

@@chrisbarber639 Hi Chris, thanks for answering. I solved it in this way: I copied the territory table (copy as reference) and deleted details that I didn't need in the first table. in this way I could connect the sales table and territory. then I connected territory table with the one with details. it works :)

[@joyo2122]

pro