Users’s don’t know, or care, how you store data in the database; nor should they have to.
It’s common today for a database front end to supply search boxes for each (potentially useful) database column:

In this case, Supplier Name and Supplier Number really are the same thing – they identify the supplier. These can be merged in one search criteria filter: Supplier.
Same with City and Province/State – they’re both an Address. The user doesn’t care that we stored them in 2 separate columns in the database; she just wants to search by the supplier’s address. So we merge them as well, giving us:

Users don’t think of addresses as seven different fields:
Street1: Johnson Janatorial Supplies
Street2: 17 Piney Park Road
Street3: Unit Nº34
City: Ashville
State: NC
Zip Code: 28806
Country: USA
They think of them as a single block:
Johnson Janatorial Supplies
17 Piney Park Road
Unit Nº34
Ashville, NC 28806
USA
It’s not the user’s problem if an address is stored over multiple database columns. It’s the job of the computer to search them all as though it were one block of text.
Taking these ideas further, people have seen how powerful Vista’s new single search box is.

The user doesn’t need to know where a file is, or it’s name; she only needs to remember something. After that, it’s the job of the computer to find it.
This also works for business applications. The user doesn’t care if the supplier has an ID, a Name, a Short Name, an Accounting System Name, a Legal Name, and a Display Name. She only cares that she saw a Johnson:

It’s much more powerful to find something when you can just keep trying words – adding them as you narrow the results.
Detailed search criteria can be useful; mostly when you’re picking between existing groupings (e.g. active/inactive, gold/silver/bronze, cash/credit, customer/employee, etc), or trying to generate a report.
But for blind searching, a single box that does all the work for you is much more powerful.