Question Regarding Ordering of Objects and Suggestion

Neil DavidsonNeil Davidson Posts: 281
edited November 25, 2005 9:36AM in SQL Compare Previous Versions
Grant,

Thanks for the feedback. We're currently working on a better way of ordering
the dependencies but it's a significant piece of work and is likely to take
a few weeks.

We're also going to be working on the error handling.

Yours,

Neil Davidson
"Grant Case" <grant.case@norspambpleasec.com> wrote in message
news:GaJw57jsDHA.1488@server53...
> We are currently using SQL Compare 3.0 and I can say without a doubt it is
a
> much FASTER and better product!
>
> I had a question though as to the logic determining ordering of objects
> creation. It seems as though when synchronizing objects that SQL Compare
is
> just synchronizing objects in no particular order instead of determining
> dependancy. For instance when one view calls another view the "called
view"
> should be first in the script followed by the caller. However, it just
> looks as though no determination is made when creating the output script.
> This creates headaches for me because I have dependancies between udfs,
> views and stored procedures.
>
> Second, it would be nice if the object that errored in a synchronization
> done within SQL Compare was also identified. Instead of the error:
> Invalid Object Name BFGWEB.dbo.PBFG_Accounts
> The error would show
> Error in Creation of Object VW_Borrower_Accounts - Invalid Object Name
> BFGWEB.dbo.PBFG_Accounts
>
> These two things would help me TREMENDOUSLY when trying to troubleshoot
down
> sync problems and building new databases from our production schema for
our
> test environment. Thanks.
>
> Grant
>
>
- Neil Davidson
Red Gate Software Ltd

Comments

  • Brian DonahueBrian Donahue Posts: 6,590 Bronze 1
    SQL Compare stopped using Microsoft SQL Server's sysdepends table to get dependency information about SQL Server objects because it is only updated at object creation time and doesn't track dependency changes as objects are modified.

    We now have our own custom dependency generator which runs as databases are being scripted (registered). This adds a bit more time to the database registration process, but the accuracy has improved quite dramatically!
This discussion has been closed.