Are incomplete records dangerous?

Dwight

Registered User.
Local time
Today, 21:22
Joined
Mar 17, 2003
Messages
168
I am hoping to avoid future problems. My question involves incomplete entries into tables. Will they have negative future consequences and if so how do we deal with them?

Here’s the situation: Each of our clients has an asset allocation to fit their needs. One client may require 40% equities, 60% bonds and another 20/80 or whatever. I have a form that allows the user to make a selection from a combo box to select the asset class (equity, bond, etc) and another field so they can enter the percentage. This is stored in a table with an autonumber PK called AllocationGuidelineID, a client ID field, asset class field, and percentage field.

Each entry is a separate record in the table i.e. the 40/60 split above would be two separate and unique entries into the table, one for the client’s bond and one for the equity allocation. I think this is the correct method but let me know if it is not.

Here’s the question: If a user selects an item from the combo box for a client and then changes their mind or realizes they are making a mistake the table will have an entry but it will be incomplete. The only populated fields would be the PK because it is an autonumber and the Client ID. The asset class and percentage fields would be blank. Is this a problem?

Do database administrators set aside time to periodically clean these entries out or do they leave them be because they are harmless?

I realize I could set a required field on the percentage field and this would not let the record be created but I would like to assume for this learning example that the empty record is created.

Hope I have not muddled this for you.

Kind regards, Dwight
 
Dwight,

If you are aware of Null data, then it does not have to
be a problem. If you are not aware of it, or don't expect
it then it can lead to severe problems.

Consider a survey form that people fill out. There is a
field (sex = M/F) on the form. 1,000 people fill out the
survey. Fifty people don't want to fill in the field stating
their sex.

If you encounter this data for the first time, get the total
where sex = M (450) and sex = F (500), you might assume that
950 people participated. If you then based your percentages
on the 950 sample size, your stats would be skewed.

Just some thoughts...

Wayne
 
Thanks for the advice. The consensus seems to be to prevent the problem before it occurs. I can do that.
 

Users who are viewing this thread

Back
Top Bottom