The ultimate Fabric, Power BI, SQL, and AI community-led learning event. Save €200 with code FABCOMM.
Get registeredCompete to become Power BI Data Viz World Champion! First round ends August 18th. Get started.
I am currently working on a project where I primarily query data from a single SQL database using Power BI. I understand the significance of using a star schema for data modeling, especially when dealing with multiple data sources. However, I am curious about its importance and benefits in scenarios where the data is primarily sourced from a single SQL database.
My question is: Is it advantageous or necessary to build a data model using a star schema in Power BI when all the data is being queried from a single SQL database?.
I have a hypothesis that building a star schema might be more relevant when importing and integrating data from various sources rather than a single database. But I am seeking insights on:
I would greatly appreciate any insights, experiences, or advice from the community regarding this. Understanding these aspects will significantly help in making informed decisions for my current and future projects.
Thank you in advance for your time.
Hi @HamidBee ,
If your data model is relatively simple and performance is not currently an issue, you may continue with a single-table approach but keep an eye on performance as your dataset grows.
Without a star schema, you might face difficulties in scaling your model as new data sources are added or as reporting requirements become more complex.
Best Regards,
Xianda Tang
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.
A single SQL Database can consist of 100s of tables and views - what do you mean?
Or do you mean you work only with ONE table? Where should the star schema come from then? 😉
Let's asusme we have a SQL db which consists of 10 tables joined together in a star schema. We decide to build a power bi report with data based on these tables. We use DirectQuery, in this case would we still need to build a data model (assuming this is the only source). I'm guessing the answer is no, but I just need to be sure.