Forum Discussion
Understanding datasets storage
Hi everyone,
I'm working with my team, and we use multiple clouds and xlsx files to do our reports. I still don't know why sometimes a PBIX could be 1.2 GB and other times only 1 MB; what's the difference? How to make all dashboards less heavy?
Also, what's uploaded in Dataset terms when I publish from my desktop to PowerBI Service? Is it the data? The connection? (we are using Databricks)
Regards
Hi julsr ,
The size of a Power BI PBIX file can vary greatly depending on several factors:
1.Data Volume 2.Model Complexity 3.Visualizations
To reduce the size of your Power BI dashboards, you could consider the following strategies:
1.Optimize Data Model. 2.Reduce Data Volume. 3.Use Calculated Columns Wisely.
Please refer: How to Reduce the Size of Power BI PBIX File by 70% (xavor.com)How to Reduce the Size of Power BI file in a few Steps - RADACAD
When you publish a Power BI Desktop file to the Power BI service, you’re essentially uploading the data in the model to your Power BI workspace. This includes any reports you created in Report view. If you republish your Power BI Desktop file, the dataset in your Power BI site is replaced with the updated dataset from the Power BI Desktop file. Any changes you make to the report in Power BI won’t be saved back to the original Power BI Desktop file.
In terms of Databricks, if you’re using DirectQuery or Live Connection mode to connect to Databricks from Power BI Desktop, then only metadata (like table schema) and not actual data is uploaded when you publish to Power BI Service. However, if you’re using Import mode where data is imported into Power BI Desktop from Databricks, then that imported data is uploaded when you publish to Power BI Service.
Best Regards,
Neeko Tang
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.
3 Replies
Hi julsr ,
The size of a Power BI PBIX file can vary greatly depending on several factors:
1.Data Volume 2.Model Complexity 3.Visualizations
To reduce the size of your Power BI dashboards, you could consider the following strategies:
1.Optimize Data Model. 2.Reduce Data Volume. 3.Use Calculated Columns Wisely.
Please refer: How to Reduce the Size of Power BI PBIX File by 70% (xavor.com)How to Reduce the Size of Power BI file in a few Steps - RADACAD
When you publish a Power BI Desktop file to the Power BI service, you’re essentially uploading the data in the model to your Power BI workspace. This includes any reports you created in Report view. If you republish your Power BI Desktop file, the dataset in your Power BI site is replaced with the updated dataset from the Power BI Desktop file. Any changes you make to the report in Power BI won’t be saved back to the original Power BI Desktop file.
In terms of Databricks, if you’re using DirectQuery or Live Connection mode to connect to Databricks from Power BI Desktop, then only metadata (like table schema) and not actual data is uploaded when you publish to Power BI Service. However, if you’re using Import mode where data is imported into Power BI Desktop from Databricks, then that imported data is uploaded when you publish to Power BI Service.
Best Regards,
Neeko Tang
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.
If you use Import mode, data will be uploaded to Service.
Using DQ mode only connection information will be uploaded.
Thats what I have seen so far on this topics.
Regards.Hi julsr ,
The Investigation of Data Storage Size usually requires couple of things to consider
1. If it is an import mode data storage that you are using, the entire data will be uploaded along with the dataset into Power BI Service. In DirectQuery Mode, only the metadata will be published to Power BI Service.
2. If it is import mode and the data is also less, then you might have to check other parameters like
- is your Data Model as appropriate and optimized?
- Are there any cross joins happening and under cetain conditions. This you should do since you are having this issue only at certain times
3. The Other thing you should consider checking is the Visualizations used, Some visualization take up additional capacity.
4. The Data that is uploaded into Power BI is usually highly compressed. This compression could be as much as 90%. But for this compression to work best the data model that you are using in the report should be optimized. Star Schema model is the recommended data model in Power BI
Regards,
