I’ve seen the suggestion below before to avoid ORMS on the read side. This may be appropriate on occasion when you have complex queries or object graphs but as a universal rule (eg. “Don’t use ORM on the read side”) this gets a 👎🏻 from me.
Every modern ORM can select specific columns without “loading the earth”. Entity Framework for example can select columns into either an anonymous type or a strongly typed DTO class.
Developer ignorance about how to do this does not mean it’s not possible. You give up a lot of maintainability benefits if you never use an ORM for the read side of your model.
My initial share of the original post on LinkedIn
