Declaring user-defined data type differences/usage

[kamfaima]

What's the difference between:

• Defining a UDDT as an array of objects, then defining a parameter to be that type; and
• Defining a UDDT as an object, then defining a parameter to be an array of that type

i.e.

param myParam myType

type myType = {
  name: string
  location: string
}[]

vs

param myParam myType[]

type myType = {
  name: string
  location: string
}

In my testing, I've noticed no difference in outcome so I suspect it's just a personal preference but I just wanted to be sure. I've also noticed that MS seem to prefer the first syntax which to me, I can't immediately tell if a param is an array or an object—instead, I have to first find the type definition.

[johnlokerse]

Hi @kamfaima, there is no (technical) difference between those two examples. It's just a syntactic difference...

If we take a look at Azure Verified Modules (AVM - Key Vault) you also see this, but the focus is more on the name of the parameter. For example, in AVM a type that is an array is written in plural (e.g. accessPolicies or secrets) and other types such as string, int, bool are written in singular (e.g. enableRbacAuthorization or name).

So it depends on your own Bicep code convention how to deal with this. Both options are fine though!

[kamfaima]

Ok, thanks for confirming that there is no technical difference—I guess it just comes down to what your preference is.