Our accounting software uses SQL (on a different server) as a backend,
and I'm allowed read-only access to it.  I pull a lot of data from it
related to our employees, contacts and company projects using a DTS
package that drops tables locally and rewrites them from the accounting
db.

However, I would have the tables set up differently or certain columns
added to certain tables but because I currently drop table and rewrite
I can't do it that way...all of "my" data must added via views/joins,
etc.

I would like to establish better/strong naming conventions and
reorganize everything...fortunately most of the tables I pull from have
a last modified datestamp.  So, is there a big performance/load issue
with overwriting only certain fields based on modified date and then
checking for records that don't yet exist in my DB and doing separate
inserts?

One issue with this technique I've thought of is that other people
looking at the DB may not realize these certain fields could be
overwritten any night by the DTS package.  Currently the only person
who might be browsing the DB natively is my boss, and rarely, so maybe
it doesn't matter?

random select a row based a weight column.
Timeout expire
Optimize this query
How can I insert a new field between existing field?
store results over network