I’ve often used actual SQL, I would here too except I wanted to give Ecto a try and I seem to quite like it, though I do have to fragment stuff up a lot due to Ecto limitations, so I’ve not seen that it was required by poorly made ORM’s (thankfully Ecto is comparatively not poorly made ^.^), I’ve tried to follow SQL tenents, even with my Ecto work (which is why almost none of my models have an ‘id’ field, well they often do, but not as Ecto would expect it to be so I have to @primacy_key false quite a whole lot).
