Re: Data Model - Please Comment on my Approach

>>My team has a list of postcodes in an access database. The codes
start from 1011AA. A table stores the following Social Demographic information is stored about each Six Character (1011AA) postcode: ... <<

I would prefer to have the data from which this summary was built (one row per household with number of person, income, etc.), but if this is how the data is collected, then you have to use it. While a column with a larger number of columns is often a bad sign, but it is not wrong.

>> This is how I'm thinking of going about it.
 

1. Create a table that has all the postcodes:

CREATE TABLE PostCodes
(postal_code_id INTEGER IDENTITY PRIMARY KEY,  postal_code CHAR(6) NOT NULL);

Why do have an identity column?? It is not a key, and it allows duplicate postal codes!! The rest of your design is a meta-data model, with column descriptions. You will wind up with monster joins.  Sorry, but this is not a good idea at all. The original model is correct. Received on Sat May 24 2003 - 00:01:03 CEST