Re: A philosophical newbie issue: catch redundant errors via relationships or programmically?

From: Marshall <marshall.spight_at_gmail.com>
Date: Sat, 29 Dec 2007 14:42:15 -0800 (PST)
Message-ID: <a57ff492-91c0-4360-8de7-26ad3a2ea855_at_b40g2000prf.googlegroups.com>


On Dec 29, 6:51 am, raylopez99 <raylope..._at_yahoo.com> wrote:
>
> Everything worked fine for this CD extension, I set in up in no time,
> but I noticed that when a user enters data, sometimes the same
> combination of Bank Account number and CD identifier (CD ID) will be
> entered by mistake, which under the below architecture will create a
> new record. I don't want this.
>
> I have various ideas on how to fix this programmically (such as a
> simple lookup table so when the combination of particular Bank Account
> and CD ID are found, a warning box will tell the user that this record
> already exists).
>
> However, I can't, under the below architecture, seem to do this via
> "relationships" in Access.

Hi,

Since you asked in a theory newsgroup, you're likely to get a theory-based answer. The short, theory based answer is that checking for this programmatically is not the right way to go. Reasons include that there may, over time, be various places the data can be entered, and thus *every* program written, at every place such data is entered, will have to enforce this constraint, and you will get data corruption if there are any places you miss, or any bugs in any of the implementations you come up with. If you use the DBMS to enforce the constraint, this will not be the case.

If you need help with Access-specific issues, you'll do best to ask in an Access-specific newsgroup. I would recommend you mistrust any Access-specific answers you get in a theory newsgroup, and I would also advise you to mistrust any theory advice you get in an an Access-specific newsgroup.

HTH Marshall Received on Sat Dec 29 2007 - 23:42:15 CET

Original text of this message