Re: Identifying candidate keys and primary keys

From: Bernard Peek <>
Date: Wed, 16 Apr 2008 22:07:18 +0100
Message-ID: <>

David Cressey wrote:

> No wonder you are confused by your assignment. It sounds like the people
> who gave you the assignment did not describe the subject matter in
> sufficient detail for you to be able to determine the candidate keys. Were
> they relying on you to use general knowledge of how hospitals work to
> determine candidate keys? Were they relying on you to do online searches to
> find out which combinations of columns are unique?
> Or (shudder) did they even understand the implications of the assignment
> they were giving you?

I would expect someone returning an assignment to document any assumptions they make. The students shouldn't be marked on the depth of their knowledge of hospital administration systems.

The first commandment of data-analysis is "Know thy data." That's not possible in a situation where the analyst can't ask the business questions. The only option is to make the best of a less than ideal situation. Document what assumptions have to be made. Document what information is missing.

But also bear in mind that the same assignment is being done by a whole class. It may have been done by several classes in earlier years. Nobody expects a student to produce a perfect data model ready to be implemented in a hospital. Instead they will be assessed on how well they do compared with other students in the class, and against previous year's students.

There's a risk of paralysis by analysis. A candidate could spend a year studying hospital administration and will get a perfect answer, but it will not be ready in time to be marked. Better a good answer delivered on time than a perfect answer never delivered.

Received on Wed Apr 16 2008 - 23:07:18 CEST

Original text of this message