Is delayed data load possible (or is my concern for performance unwarranted)?

[wittr9876]

Greetings.

I have a report in which this is the desired action flow:

1. User selects a person from a dropdown list.

2. Based on this selection, a table displays training records associated with the selected person.

I have a working copy of this report in PowerBI, BUT I have a concern for performance. I am thinking that PowerBI pulls in ALL of the data (all training records, for all persons). Since no person is initially selected, there is no data displayed in the table, BUT I think it's all been pulled into memory. Then, when a person is selected, the appropriate data is displayed.

There is potentially a LOT of data that would get initially pulled in to this report (eventually for thousands of persons.). Never does there need to be more than one person's data in view. Is it possible to perform a delayed load, i.e., load NONE of the training data from SQL Azure until a person has been selected, THEN query SQL Azure for the appropriate training records? And, if a different person is selected, re-perform the SQL Azure query?

Or, is my concern for performance unwarranted, i.e., is a one-time load of a large amount of data, even if most of it's never accessed, preferable to repeated access of SQL Azure for small quantities of data?

Thanks,
Randy

[v-yuezhe-msft · Accepted solution]

@wittr9876,

You can use DirectQuery mode to connect to SQL Azure database in Power BI Desktop. With this mode, no data is imported into the Power BI store. Instead, upon building a visual within Power BI Desktop, queries will be sent to the underlying data source to retrieve the necessary data. The time then taken to refresh the visual will depend on the performance of the underlying data source.

Reference:
https://docs.microsoft.com/en-us/power-bi/desktop-use-directquery
https://docs.microsoft.com/en-us/power-bi/desktop-directquery-about

Regards,
Lydia