Forum Discussion
large datasets - premium
Hi,
Was reading about the large data models feature as below.
Looking at feasability of premium with large data models.
https://powerbi.microsoft.com/en-us/blog/large-models-in-power-bi-premium-public-preview/
As I understand it, have to keep the .pbix file below 1gb initally when upload to service but can use incremental refresh to allow dataset to grow to several hundred gb.
Have a few of questions
would wangt to use import mode for instant speed as direct query not really an option due to length of time would take to return.
1) If need to download the .pbix file from service to amend on in desktopat later stage (e.g after initial large load) has been done, would it allow this and how large would we expect it to be? Assuming the data stored separatelly in the service to the .pbox file but when look at .pboxi in desktop generally has all the data.
2) If say uncompressed 140GB Oracle what sort of compression ratio coudl we expect - have read separately 7 to 10x of uncompressed sql server data
3) How quick could we expect naviagtion to be on low level data approx 1 billion rows would rather avoid maintenance overhead of creating aggregations
Any thoughts of anyone using power bi with large datasets
Thanks
2 Replies
- v-xicaiCommunity Support
Hi po ,
For the question1, you can download the pbix from following option. From the article, we know that the Pro or Free users can publish 1 GB pbix files from Power BI Desktop to Power BI Service, while the users in Premium capacity can publish 12 GB pbix files. So the data model size 1 GB and 12 GB is the max download size limitation.
For the question2, you can refer to the article,
However compression rate isn’t always like that, it depends on type of data set, data type of columns and some other facts. The bottom line in this section is that, if you have a data set that is more than 10GB and won’t grow significantly, then consider loading that into Power BI to check how compressed it would be, maybe you get it under 10GB limitation and then you are good to go. If that is not the case then continue reading.
For the question3, it depends on the data source and aggregation, even the network speed.
You may improve the performance of the large dataset, see Power BI performance best practices . And you may use Incremental refresh for your data source, the Incremental refreshes only data that has changed. Since incremental refresh is Premium only feature, so you may need to purchase Power BI Premium capacity for your organization, see How to purchase Power BI Premium. Incremental refresh policies are defined in Power BI Desktop and applied when published to the Power BI service. See more detail steps: Incremental refresh in Power BI Premium .
Best Regards,
Amy
Community Support Team _ Amy
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.
- poPost Prodigy
Hi,
Thanks for reply.
When you mention cannot download .pbix over 10GB just thinking of the practicalities of how change a report once import dataset has grown very large.
1) Upload .pbix to service under 10gb
2) Run historic load in service and incremental refresh such that had imported say 100GB which sounds like possible via
https://powerbi.microsoft.com/en-us/blog/large-models-in-power-bi-premium-public-preview/
3) Then need to make change to report - how can this be achieved?
Would we need to change the initial report uploaded in step1 and then repeat step2 which would be less than ideal?
Thanks