How can you transform a relational erwin DM model into a NoSQL model.

This will pop up 3 pre-set options:

1 - A heavily Normalized model, referencing where possible

2 - A heavily Denormalized model, duplicating data and embedding where possible

3 - A Custom option, allowing the end user to choose which parts should be embedded or nested

Normalized model example result (most relational)

Denormalized model example result (data duplication and embedding where possible)

The transformation concerns itself with how to manage any relationships found within the relational model. We have rules for some relationships that meet criteria and we will handle those. Your input may be required as to how to handle the remaining.

This is where you may wish to use the ‘Custom’ transformation option. Once this option is selected, we present some questions that you should think about when choosing how to handle the relationships that are left. 

All relationships can be expanded and the list can be searched. Our example only asks us for input on one outstanding relationship. The other 2 had potentially nullable parents and the tool has handled these.

It presents us with a pictorial representation of the data object we are asking about and shows the 0:* relationship to the parent, that is in question at this point.

The user can then choose to leave this as Embedded (default), or use the toggle to make it a reference. When they select the Transform option, the decisions made will be used to generate their NoSQL model representation.

Users can produce different models, by investigating using the different options for the discovered relationships in the transformation wizard.

The transformed model would show up on the Home page with a green background and you will also receive an email confirming once the action has completed. The user will also receive and in-tool notification about the transformed model.

 Once this is complete you can open the new model from the home page.