I have to admit, it's been almost two decades in web but this is the first I'm getting into a project where we're considering actually duplicating some data.
And our use case perfectly fits what you described as ideal for this - heavy read opreations. We actually have just a single import that creates the graph db structure and everything else is read after that.
Before this I couldn't think of a situation where I would have duplicated data, but looks like I was working with a limited scope of apps. :) Thanks for the mention Saurabh!
Normalization and denormalization are two key concepts in database design, each serving a specific purpose. Great write-up.
Thanks Fernando
I have to admit, it's been almost two decades in web but this is the first I'm getting into a project where we're considering actually duplicating some data.
And our use case perfectly fits what you described as ideal for this - heavy read opreations. We actually have just a single import that creates the graph db structure and everything else is read after that.
Before this I couldn't think of a situation where I would have duplicated data, but looks like I was working with a limited scope of apps. :) Thanks for the mention Saurabh!
I often look at the decision as being about integrity vs. speed.
Normalization: Optimizes for data integrity and minimizes redundancy, ideal for transactional systems.
Denormalization: Optimizes for read performance by reducing joins, ideal for analytical queries.
Simply put, Saurabh!
Thanks for mentioning my article on EDA!