Question

I am saddled with an ERP database which lacks any foreign keys and therefore lacks referential integrity.

I am writing a script to check most of the major tables in our database for ophaned records.

For example, in this case I’m working with our Sales tables.

SLCDPM - Customer Master Table
SOMAST - Sales Order Master
SOITEM - Sales Order Items
SORELS - Sales Order Releases

Basically, for these (and a whole bunch of other tables) I need to check to see if there are records in the SORELS that don’t appear in any table above it. Then take SOITEM and check above it. Etc.

I started writing scripts, but the number of lines gets kind of ridiculous. Here is where I started with just these 4 tables.

select 'Sales Order Master',* from somast where fcustno not in (select fcustno from slcdpm where ftype <> 'P')
SELECT 'Sales Order Item',* FROM soitem WHERE fsono NOT IN (SELECT fsono FROM somast)
select 'Sales Order Release',* from sorels where (fsono+finumber) not in (select (fsono+finumber) from SOITEM)

The reason I stopped was that I just realized that SORELS (the bottom table) only checks the table before it, not all of the tables before it.

Anyone know of a script I can use to make this more automated or a better way to do it?

Answer 1

And people sell database junk like this; it always amazes me what I see in commercial products.

This is genuine case for dynamic sql and a cursor, I think. This is exactly the kind of one-time administrative function that is reason why these techniques exist (they weren’t really meant for production code mostly, but for administrative tasks).

I’d create a table showing each table and the table I think it should have a foriegn key to. (You may even be able to populate this from the system tables if they at least had a good naming convention).

Then I would use a cursor to go through the table and create the sql dymanically for looking to see if the FK table has orphaned records.

It’s still alot of code but at least you didn’t have to write it all.