Good article. It’s just that “denormalized” is a trigger word for me. I don’t consider dimensional designs to be “denormalized”. First of all, even 1NF is “normalized”. Secondly, in a good dimensional design, fact tables are 3NF and dimension tables are in 2NF. Better to think of dimensional designs to be redesigned or reorganized from their transactional design.
Thank you so much, Donald, for your comment. We hadn’t considered this perspective before, but you make a great point. Based on your view, it’s more accurate to highlight that the levels of normalisation differ, rather than simply labelling it as denormalised. That really makes sense!
Thanks for writing article, it’s a lot easier to critique than write. I’ll write up my own explanation someday. Been reading a lot of C. J. Date lately also. Have a good weekend.
Hi Sunday,
Thank you for your comment,
Sure we try our best to help, could you please elaborate the other two data models ? Are they database schema ?
Is your end goal to have a JSON schema ?
Thanks for this,please I need more info about unification of 2 data models and JSON in Schema.
Good article. It’s just that “denormalized” is a trigger word for me. I don’t consider dimensional designs to be “denormalized”. First of all, even 1NF is “normalized”. Secondly, in a good dimensional design, fact tables are 3NF and dimension tables are in 2NF. Better to think of dimensional designs to be redesigned or reorganized from their transactional design.
Thank you so much, Donald, for your comment. We hadn’t considered this perspective before, but you make a great point. Based on your view, it’s more accurate to highlight that the levels of normalisation differ, rather than simply labelling it as denormalised. That really makes sense!
Thanks for writing article, it’s a lot easier to critique than write. I’ll write up my own explanation someday. Been reading a lot of C. J. Date lately also. Have a good weekend.