Kurt Buhler is going to raise my blood pressure this morning:
Most Power BI models are case-insensitive, meaning that “Bonk” is the same as “BONK”. However, Power BI data models can also be created as case-sensitive if you create a Direct Lake model in Fabric, or create a new model with external tools and enter a case-sensitive collation property. Two otherwise identical models which differ only in this case-sensitivity may produce different results, even though they’re using the same data, DAX, relationships, and tables.
It’s useful to know how case-sensitivity affects your model and its query results. You should also be able to identify and validate whether your model is case-sensitive. This is particularly important in the following scenarios:
Read on for those scenarios and how you can fix the problem of case sensitivity. My official stance on case sensitivity, by the way, is that applications should be case-insensitive on input but retain casing on output, so “dog” = “Dog” = “DOG” for sorting and querying, but if I saved “Dog” then that’s what should display.