Our customer SEPE, has an old app that mantain some tables in an Oracle Database supported by a Solaris platform. This tables have some columns defined with attributes "Not null" and "Default value" for some Numeric, Alphanumeric and Date fields. The app doesn't validate if the user enters a value for these fields or not, and is Oracle who returns the error if the user doesn't fill some of these fields.
Now, developers are changing de architecture of some parts of this app and the Oracle tables, from now, will be replicated with Event Replicator from an Adabas database residing on the IBM Z/OS mainframe platform. They don't want to change the interactive part of the app by the moment.
The problem arise because ARF-ART don't implement the functionality of columns with attributes Not Null and Default Value that Oracle Database use. When you load a table with Populate function you don't have the possibility of define these atributes for a column. The same problem happen if you instruct ART to use Sql Loader for the loading.
If you make an Alter Table after the creation of the table to define the attributes, ART produced an error when you insert/update data in the table because fields filled with null value comes from Adabas for the fields don't filled by the app on the origin Adabas database.
SEPE would consider valuable if SAG R&D thinks on the possibility of improve the product contemplating this funtionality when replicate data to Oracle database tables. Nowadays, SEPE bypassed the problem doing an ALTER of the tables, defining these columns with NULLABLE and DEFAULT VALUE attributes, after the creation of the tables. The problem is that the time of Populate is incrased a lot, because they can't use the Oracle SQL Loader funtionality for massive loads.
Thanks in advance
|Created on Brainstorm||10.11.2016 12:44 pm|