Ron, I'm afraid this not gonna happen - too many changes in the code required to impelement this. Having more than one or two key columns is not a good idea from any point of view. Just my two cents.
50033310/6/2005
OK, I can accept that your application will not handle
this, however, seeing you have such a strong view on
key fields I'm interested in your comments on alternative
ways of dealing with situations as below: Each country has a regional office with customer numbers
SE_CUSTOMER_ID, generated at the country (not global) level.
The customer numbers may overlap, so we need SE_COUNTRY_CODE. Each customer may have a number of site locations (stores)
represented by SE_CUSTOMER_SITE_ID.
Each Site may have one or more processing units represented
as SE_EQUIPMENT_ID. Each processing unit has one or more
equipment components represented as SE_COMPONENT_ID. These five fields have identified the component, we need
one more to identify what we are trying to capture.
SE_ERROR_NUM. These six fields (marked ) now make up the primary key,
we are basically monitoring the status of each piece of
equipment SE_STATUS from a single WW support centre SE_COUNTRY_CODE
SE_CUSTOMER_ID
SE_CUSTOMER_SITE_ID
SE_EQUIPMENT_ID
SE_COMPONENT_ID
*SE_ERROR_NUM
SE_STATUS
SE_STATUS_STAMP
SE_ACKNOWLEDGED
.... Besides concatenating fields, (which would make reporting difficult)
I can't see an obviuos solution I'd appreciate your thoughts Thanks, Ron