Generic Data Reconciliation

kduke · Post by **kduke** » Sat Mar 05, 2011 1:33 pm

All this is documented on my web site, in a blog and with several posts on this web site.

ray.wurlod · Post by **ray.wurlod** » Sun Mar 06, 2011 1:36 am

It wasn't just reconciliation - it was also transfer the tables' data and, indeed, create the target tables. Then reconciliation to prove that it all worked.

kduke · Post by **kduke** » Sun Mar 06, 2011 7:01 pm

Will you are on your own for the rest.

ray.wurlod · Post by **ray.wurlod** » Sun Mar 06, 2011 7:05 pm

Yes, I was. Notice the "resolved" status, though.

:D

kduke · Post by **kduke** » Tue Mar 08, 2011 4:53 pm

Try to help a buddy out and this is all I get. Need a quick whip from Craig about now.

chulett · Post by **chulett** » Tue Mar 08, 2011 5:02 pm

abc123 · Post by **abc123** » Thu Sep 08, 2011 3:03 pm

Ray, can you tell me the steps/method you used for 3(a) and 3(b)?

Thanks.

ray.wurlod · Post by **ray.wurlod** » Thu Sep 08, 2011 5:02 pm

I've moved on from that client since then. From memory, 3(a) was simply SELECT COUNT(*) FROM TABLE for both source and target, add artificial key ('X'), join and compare.

3(b) was very similar only it was SELECT * FROM TABLE ORDER BY keycolumn (with RCP enabled) and run through Checksum stages before the remainder. To be generic the SQL was built in an upstream server job that queried the system tables to get they key column name(s), and was then passed as a parameter to the comparison job.