Hi,

I have created a dotnet windows application (dotnet 3.5, visual studio 2008 editor) and sql server 2008 database as backend.

Whenever network is available my application connects to server database
and when network is not available then application connects to local database.


But next time when application is opened then data and also schema should be synchronized between local database and server database.

So please tell me how to do this.

Also primary key columns are set as auto increment in each table.

Thanks

AND the schema?

Doesn't this just sound like a very bad idea to you?

Hi Pras

You can use database replication to sync databases, but there can occur data conflicts. I have to agree with Brett, it's not good idea.

Regards
Mike

Specifically you'll want to use Merge Replication.
I don't know how it will handle the out-of-sync identity key values.

And as OFTEN as that happens...it never gets old

You should change the data type of the PK's to GUIDs.

God Forbid

Why? It's the only case I'd ever use GUIDs.

For merging datasets? GUIDs are the best solution.

Your better of Architecting a better solution

Heavens to murgatroid. I use GUID values and UTC dates/times for any database that requires merge replication. Anything else would beat me bloody!

-PatP

Any problems with Duplicates?

Using carefully contrived lab procedures I can sometimes generate duplicate GUID values. There are a couple of ways that I can do it repeatably, but I can't imagine those happening in the real world.

Because the GUID is a fixed size, it is guaranteed to have duplicates occur naturally in due time... The most conservative estimate I've seen for an expected collision is over 1800 years, and most estimates are in the tens of millions or more years. Even at 1800 years I don't expect any code I write or even the data associated with that code to be around anymore, and if it is around then I'm pretty comfortable that someone will have a convenient way to deal with the problem by then!

-PatP