Oved,
if you are using transactional replication with immediate
updating subscribers, this can't occur as the
subscriber's change isn't delayey - it is run as a
distributed transaction with the publisher. If however
you have queued updating subscribers or are using merge
replication, then the way to avoid this is to partition
your data. EG you can have an identity column and assign
different ranges to the publisher and subscriber.
Alternatively you can have an extra column which
designates the source of the data and then use a multi-
column primary key.
HTH,
Paul Ibison

(recommended sql server 2000 replication book:
http://www.nwsu.com/0974973602p.html)

you must use a partitioning scheme where each subscriber/publisher has its
own pool of identity ranges to use. Either that or add a column to the table
with a default indicating what the server name is.

For more info search on Managing Identity Values in Books on line

--
Hilary Cotter
Looking for a SQL Server replication book?
http://www.nwsu.com/0974973602.html


[quoted text, click to view]

Hello there

I have two databases on far place

In some hour of the day i replatate them

But if i add line on the publisher with primary key
and on the subscriber i add also another line with the same primary key on
the subscriber there will be two lines with the same primary key

What i need to do in this case?