Getting the structure over is easy and I would do this through Script not
DTS. (More control)
Then you need to get your data. You are going to have to take it in order
of it's relational integrity.

Example

Authors
Titleauthors
Titles

If the Authors table had 10,000,000 records, the titleautors table
11,000,000 and the Titles table 9,900,000 then If I only wanted the top 1000
authors (ordered by name) I would then only need books they have
written,helped to write and their titles.

It will help to keep things relationally tight.

--

----------------------------
Allan Mitchell (Microsoft SQL Server MVP)
MCSE,MCDBA
www.SQLDTS.com
I support PASS - the definitive, global community
for SQL Server professionals - http://www.sqlpass.org



"vgsatwork" <vgsatwork (AT) yahoo (DOT) com> wrote

In article <#cBR0jhZDHA.2620 (AT) TK2MSFTNGP09 (DOT) phx.gbl>, Allan Mitchell
<allan (AT) no-spam (DOT) sqldts.com> writes
<snip>

I wouldn't bother about the order, because I would do the scripting in
two parts. First script all objects, excluding indexes and keys.

Next use the Wizard to create a package that exports all tables from
server a to server B.

I would then open the wizard package and change the source from a table
to query select top 1000 from table. You could actually write some
simple ActiveX Script to do this if there are a lot of DataPump tasks (1
per table).

Run the package.

Then create a second script of indexes and keys and run that.

Creating the two scripts is quite easy in Enterprise Manager, just check
or uncheck the relevant options at each stage.


I have done quite a bit of migration between sort orders and collations
recently and this method has worked well. The only problem I have has so
far is forgetting to apply the second script!

--
Darren Green (SQL Server MVP)
DTS - http://www.sqldts.com