For some options, check out the article "Auto-Increment and Unique
Identifier" on page 8 of April 26, 2006 issue of Centerfield newsletter you
can find at this link:
http://www.centerfieldtechnology.com/publications/
Although it has some caveats, my favorite is the Identity column attribute.
HTH, Elvis
Celebrating 11-Years of SQL Performance Excellence on IBM i, i5/OS and
OS/400
www.centerfieldtechnology.com
-----Original Message-----
Subject: Best way to make a table of non-unique rows, unique?
Gang,
I have some tables from a vendor using OLD, OLD, OLD, non-relational
architecture which allows duplicate rows. I need to find a way to make the
rows in these tables capable of having uniqueness for a unique key, even it
the rows are "copied" and kept updated to like-type tables.
Since I can't touch the original tables because they are from a vendors
app., I was thinking of copying the rows in the original tables to new
tables with an added column wherein I could put something like a
datatimestamp to make the row unique. I'd like to use something at the
database level like a trigger or stored procedure but, if necessary, I'll
write an old fashion program to do the work.
The reason for all this is that I'm using DataPropagator to map and keep in
sync tables in a data warehouse in a separate LPAR and DataPropagator has to
build a unique key over the target tables based on whatever makes rows
unique in the source tables. Well, if the source tables have duplicates, I
have to find a way around that or DataPropagator won't work for the last 11
tables I have to map/keep in sync.
Thanks in advance,
David Odom
midrange.com
