restrict loading of dataset

[Anonymous]

Can i restrict number of rows to load in dataset by applying where clause.

Requirement is i dont want complete data to be imported in pbix but need data for 10 clients only.

Instead of applying where clause in sql end, can i do something at pbix end so it doesnt load all clients data?

Kindly help.

[Stachu · Accepted solution]

if you filter the customers within the PowerBI interface, it should apply query folding to apply the filter
https://blog.crossjoin.co.uk/2017/06/11/query-folding-and-writing-your-own-sql-queries-in-power-quer...

[Abhilash7181]

In Option 2, the entire dataset is first loaded into Power BI, and then filters are applied at the visual level. This means that while the report contains all client data, only the restricted client information appears in the visuals.

However, loading the full dataset before filtering can impact performance, especially with large datasets. A more optimized approach is to apply filtering at the source level using a SQL query before loading data into Power BI. This reduces the dataset size and improves report performance.

For example, instead of bringing in all client data and filtering later, you can use a SQL query with a WHERE condition to load only the required data:
SELECT *
FROM Clients
WHERE Region = 'West' -- Load only data for the 'West' region

This ensures that only relevant data is imported, leading to faster refresh times and improved report efficiency.

Regarding dataset size, you can check it in the Group Storage Section in Power BI Service, where it provides details on how much space your dataset occupies. Below is an image showing where you can find this information.

[Stachu · Accepted solution]

if you filter the customers within the PowerBI interface, it should apply query folding to apply the filter
https://blog.crossjoin.co.uk/2017/06/11/query-folding-and-writing-your-own-sql-queries-in-power-quer...

[Thejeswar]

Hi @Anonymous,

Yes you can filter the data set in PBIX Power Query window which you get when you click on edit Queries, just like the way you do in excel.

Click on the column drop down and filter the items that you want

[Anonymous]

Hi, 

There are two options available.Not able to understand consequences of each.

Which one should i prefer.

Option 1: Writing sql while importing my view in power bi desktop with where clause.What impact will it have to performance since it stops query folding.

Option 2: to apply select specigif id from columns populating in edit queries?Will it not pull all data from sql server and then apply filter?

[Thejeswar]

Hi @Anonymous,

I would say Option 1 is the better approach of the two specified.

In Option 1, your data will be restricted even before it is imported to your report. In this means, you can restrict the volume of data getting imported in Import Mode. In case of Direct Query, it will certainly improve performance as compared to applying filter in Power BI, as every time your query runs the query is applied on the entire table to which it is connected to

In your Oprion 2, as you say, even if you filter for specific id in the Edit Queries Window, the Filters will be applied only after the entire data is pulled in to the report

[Anonymous]

I doubt based on blog shared above.If we need to limit data in dataset, we can opt for option 2.

I am confused that if i will publish this dashboard on service, will dataset hold all clients infotrmation or only those which are filtered out.

Continuation to this can we see dataset size which we publish on workspace, individually.

[Thejeswar · Accepted solution]

In Option 2, your entire data is brought in, Only after that your filters are applied. In Other Words, Your report will have all clients information, but your visuals will get only the restricted clients information as you are applying the filters to your dataset before using in Visuals

Yes you can see the datset size in your Group Storage Section. Find below the image

Clicking on this, will display all the information related to the data sets

[Anonymous]

Thanks alot for help