Hi!
I'm pretty new to Power BI and I have a question.
I'm competant using SQL, I've been writing queries for a few years now, so I feel comfortable doing that.
I don't really understand the joins between queries/tables in the 'relationships' view in PowerBI. Of course I understand them in SQL, but I find them confusing in Power BI. You have many to one, one to many... multidirectional... If something appears more than once in a table, it can be necessary to build an intermediary table with distinct values in....
I find it all a mindfield. I haven't really put the time in in order to understand it fully, so it's my fault.
However, I recently built my first proper dashboard.
I managed to build it from a couple of SQL queries. They didn't need to talk to eachother so they didn't need any relationships.
So I avoided the relationships entirely, so all was well.
My point is this:
If you're bringing in data purely from SQL databases, would it make more sense to do all the joins in the SQL query, before the data arrives in PowerBI. That way all your columns are in the same SQL query. This means I don't need to do any joins in power bi (I don't understand them!) and also, would it be quicker computationally for all the data to be retrieved at once in a nice SQL query, rather than doing multiple SQL queries, importing them all separately and creating relationships in Power BI?
If I was bringing in data from a source other than SQL, for example a .csv, I can see that I would join in Power Bi and create relationships. I think that the only reason I would use the power BI relationships is if the data is from a source other than a SQL database.
Does that make sense? So if you're competant constructing SQL queries and are able to do all the joins and get all the data BEFORE Power BI, then why not do it then and avoid joining in PowerBI at all.
Please let me know your thoughts.
Thanks for your help
Sean
