Which practice should be used for type-safe access to properties on a query entity type?

The practice of using the property reference (EntityType#Property) for type-safe access to properties on a query entity type is highly recommended because it promotes code clarity and reduces the risk of errors. By referencing properties in this manner, developers benefit from compile-time checks, ensuring that the property names are valid within the context of the entity type. This approach also enhances code maintainability, as any renaming of properties will be reflected immediately in the references throughout the codebase, allowing for immediate feedback during the development process.

Using direct string references can lead to potential issues, such as typos not being caught until runtime, which can create challenges in debugging. Similarly, utilizing indexed column names directly lacks the type safety and clarity that property references provide, increasing the likelihood of runtime errors. Implementing wildcard searches may have its own use cases, often associated with query flexibility, but does not address the need for type-safe access to properties specifically.

Overall, employing the property reference is the best practice because it aligns with the principles of strong typing in programming, significantly enhancing reliability and readability in code related to query entity types.