Archiving Order Management Data

Question

Hello All,
During implementation phase 6 months ago, we had to convert past 2 years of sales order history which resulted in ~12million order lines to be imported in Oracle R12.1.3. Now, everyday we receive 25-30k order lines and hence the data size has increased enormously. This has resulted in huge performance issues with the custom extensions and interfaces and standard functionalities.
We can't purge old order lines due to audit and HIPPA regulations and need to archive them somehow. To add to it, we have a custom index on actual_shipment_date column of the lines table, which has timestamp as well due to unavoidable business reasons. We also have rightly sized our hardware and tuned the database to meet this large data size demand.
Therefore, any pointers on the following will be helpful:
1) How can we archive (not purge) the humongous data volume in the order lines and similar tables such as wsh_delivery_details etc. and yet be able to pull it on demand.
2) Will these archived data be somehow retrievable on the OM forms later..
Looking forward to the best practices related to data archiving in OM..
Thanks,
Rahul.

Answer 1

There is no easy answer and this is not an easy project.
You need to identify what exactly the users refer to. You mentioned Old prices. What else?
It seems that you are using the old data quite frequently. In that case, moving it off to a different table will work only if you are willing to develop a screen for them to access it.
Also try to quantify the amount of usage. When the user says "I use it a lot", he/she may mean once a day or four times a year. Get some specifics to put your arms around the true requirements. Explain to them that they may have to spend an extra minute to access the old data but by making the other screens faster, they will end up saving time.
If you must leave the data in the same table, Here is one approach
1) You create a partition based on year and quarter of the orderline creation_date. (2011-03,2011-04 etc.)
2) You have the responsibility (say OM User) that users use at the moment. You create another responsibility called OM Archive User
3) You write personalizations or custom.pll code to add a default condition to the where clause in the OM User screen. The where clause focuses only on the current and last quarter partition.
4) If the user needs to query old data, he/she uses the OM Archive user responsibility.
5) If you don't want the user to switch responsibility, you can create new functions and add them to the same responsibility. Your code will then use function names instead of responsibility name to add the where clause.
Sandeep Gandhi