Enforcing Naming Standards for Your Data Models

Customer asks:

We tried also to apply some naming conventions but I think we didn’t understand something….
Having a logical model with a table having the column X, and a glossary having the standard name as Y. Can I apply some checks based on the glossary directly in the logical or relational model?

Yes, in fact you can.

Basically how it will work is:

1. You will define a glossary, having words you approve for your object names
2. You will add one or more glossaries to your design properties
3. You will run the appropriate Design Rules when you want to do a quality check

Let’s start with creating our Glossary.

Tools > Glossary Editor.

It’s going to ask for a .glossary file. You don’t have one. So browse to a directory and type in a file name you want to use – the data modeler will create the file for you and open this dialog:

So put in your approved words, hit Save, and then Close.

Now We’ll Tell Our Design Where to Find the Glossary

Right-click on your design in the tree, and open the properties.

Now What?

The Modeler isn’t mean. It’s not going to laugh at your design. So when I create a new entity or table and try to save or close my design, it’s not going to give me an error if the names aren’t adhering to the ‘rules.’

No, instead you can ask for a quality check.

Back to the Tools menu, but this time select ‘Design Rules & Transformations’ – ‘Design Rules.’

Open the logical and/or relational set, and select the naming convention items.

Hit the ‘Apply Selected’ rules button, and get your results.

Since I had my naming standards settings for this design set to treat an underscore (-) as a ‘Separator’, it’s treating ‘LOTS_OF_WORDS’ as three separate words to check in the glossary.

So, fix my names, and run the rules until I pass everything. Or ignore the rules and wait for a design committee to frown at me.

PS – Thanks Heli!

I hadn’t used this feature in awhile, and needed to brush up on a few things. Thankfully I had Heli’s book handy 🙂

Totally worth the $30 to pick this up.