Keeping data sources in separate folder and model data only within tables that reference source

[sandra_p]

I can't recall the resource, but I read once, that it is a good practice to keep datasource tables in a separate folder, don't shape them, and then reference them and do all the data shaping and modeling in those tables that reference source tables.

I wanted to find our what are the opinions about this here 🙂  And, does anyone here uses this approach?

What could be pros and cons? Wouldn't this approach create redundant data that takes up space?

[v-yingjl]

Hi @sandra_p ,

If you've fixed the issue on your own please kindly share your solution. If the above posts help, please kindly mark it as a solution to help others find it more quickly. Thanks!

Best Regards,
Yingjie Li

[v-yingjl]

Hi @sandra_p ,

In most instances, we connect to the data source and shape data directly because what we have done in power query can all be seen in Applied steps and Advanced editor, we can restore the data source by deleting steps when we need. This sounds cumbersome but we don't need to reconnect to the data source to get table again to add memories in power bi.

Another situation is that we really need to refer to the source table, in this way, we can use 'Duplicate' and 'Reference' in power query:

• Duplicate copies a query with all the applied steps of it as a new query; an exact copy.
• Reference will create a new query which has only one step: Getting data from the original query.

• It is a good idea to uncheck the enable load on other queries to save some memory in Power BI.

You can refer this article which describes in detail the differences between them.

Best Regards,
Yingjie Li

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

[amitchandak]

@sandra_p , not very clear

do you mean

https://exceleratorbi.com.au/measure-tables-in-power-bi/

Or

https://docs.microsoft.com/en-us/power-bi/connect-data/desktop-report-lifecycle-datasets

[sandra_p]

I mean - when you connect to datasource - you get the table.

And you don't do any datashaping in that table.

You create new table, that references that first table and do you data modeling and shaping there, so your original table stays as it was retrieved from data source.