SMO is incorrectly defaulting the schema-identifier, when it shouldn't be (e.g. on CREATE PROCEDURE).

Create a database, e.g. TEST1, and create the following stored-procedure:

CREATE PROCEDURE GetServerName AS BEGIN SELECT @@SERVERNAME END  

Note: no schema identifier was provided above.

Now, in something like SSMS (or Azure Database Studio), whatever, query the stored-procedure (sp) and it clearly states that it belongs to dbo/schema_id=1.

Now, to exercise SMO, use SSMS (or whatever) and generate the SQL to CREATE the sp, and you'll get:

CREATE PROCEDURE [dbo].[GetServerName] AS BEGIN SELECT @@SERVERNAME END 

You'll notice that it's added the schema, [dbo].

I think, that this could actually be a bug. I contend that it shouldn't have added the schema.

Why? Well, apparently it makes a difference . . .

Now, create a second database, e.g. TEST2, and apply the generated sp, then compare, either with (i) Azure Data Studio's "Schema Compare" tool, or (ii) DacFx's SchemaComparison (I assume they are the same thing) and . . . it flags a difference: the SQL it generates for TEST1 doesn't have the schema, but the SQL it generates for TEST2 does have the schema.

My question is: how can I configure SMO options (for calling from C#) to only generate the schema-identifier value it was originally provided (or nothing, if that's the case)? With that, I could run DacFx/SchemaComparison and the two objects would be considered identical.

There could be a little debate as to whether the issue is SMO or DacFx/SchemaComparison, but it seems logical the problem originates in SMO.

Note: the same problem applies to views, UDFs, etc.

Reproduisez le problème avec les scripts de procédures stockées TEST1 et TEST2, puis examinez les options de scripting de SMO et le comportement de son API pour les procédures, les vues et les UDFs. Comparez les définitions générées avec Schema Compare ; le travail est terminé lorsqu’un schéma omis reste omis, afin que les objets équivalents ne soient pas signalés comme différents.