I am considering using streams for replicating data from our production
ERP system to a data warehouse. This will be a fairly straightforward
unidirectional transfer, including some minor data transformation. By
minor, I mean:
- removing columns
- adding columns
- renaming schema
The data is simple table data -- no LOBs, types, nested tables, IOTs, or
anything like that. Eventually, the data will be purged from the
production ERP system, so I additionaly want the replication to
not
apply 'delete' operations on the target database. Finally, my source
ERP system has like 800 tables, but I only need about 30 of them in the
data warehouse, so I only want to replicate a select subset of my
database. Eventually, I would like to have multiple ERP systems feeding
the same data warehouse, albeit to seperate tables and/or schemas.
