How to implement CDC in Datastage Server ?

Bilwakunj · Post by **Bilwakunj** » Thu Dec 16, 2010 9:59 am

Hi-
In past I've used CDC stage to do change data capture in PX at a table level but now I need to know how CDC could be implemented using server edition. Also, my requirement is to do CDC at the schema level, in other words I need to know what records have been inserted, updated or deleted from all transactional tables of that schema and bring that data to staging area. The count of transactional tables is close to 300 and generally total record count of each table is close to 10 million records.
If I decide to go for 8.x server , will there be any difference in implementation of CDC described above?
Thanks in advance.

ray.wurlod · Post by **ray.wurlod** » Thu Dec 16, 2010 3:23 pm

Code: Select all

TargetTable  ---->  HashedFile
                         |
                         |
                         V
SourceData ----->  Transformer  ------>  TargetTable
                         |
                         |
                         V
                    HashedFile

The two hashed file stages refer to the same hashed file. Do not use read cache or write cache. Do use lock for update.
This design will work equally well in version 7 and 8.
Bring as few rows and columns as possible into the hashed file.