Idea. Add a column or model level setting to keep the exact casing of each text value on load, for example a case sensitive option for text columns.
Why. Power BI treats text values as case insensitive and shows the casing of the first value it encounters. In a reported example a City column with PunE, Pune, pune and PUNE loads as four rows that all show PunE. This is expected engine behavior, but it surprises people building connectors and models where the original casing matters, for example codes, identifiers or names.
Suggested behavior. An opt in setting per column or per model that preserves the source casing and treats differently cased values as distinct, with clear documentation of the storage and performance impact.
Benefit. Fewer workarounds such as adding a helper column with a hash or a prefix, and more predictable results when the source data is case sensitive.
Source thread (raised by another community member) https://community.fabric.microsoft.com/discussions/power-bi-services/x/5368143