I'm looking for advice as I've got to read a million+ rows from some tables
on one server, transform and merge some of the data and write it out to a
data warehouse on another server. The servers are on the same network. I was
planning to create a scipt with one server as the source and the other as the
destination, but I'm being pushed towards having to extract the tables to
flat files on one server, FTP to the other server and then load them into the
warehouse. The reasons I'm being pushed to FTP are to minimise network
traffic and to reduce the connection time to the source db.
Has anybody got any views on how network traffic and connection times are
likely to vary between these two approaches or does anyone have any other
feedback?

Thanks
--
Darren

If you have millions of rows and the row is very wide, then it is always a
good idea to export it using bcp, zip it, ftp to target, unzip it, and bcp
it in.


"DarrenS" <DarrenS (AT) discussions (DOT) microsoft.com> wrote