Re: Selecting SIMILAR, not the same records (PROBABLE) duplicates

G Quesnel wrote:
> Now that you have explained some of this, people might be able to
> provide better options. For example, if you created a trigger or a
> procedure to check for possible duplicates, as the data is captured,
> then you would not need to join the table with itself, but rather, do a
> lookup on just one table, compared to some 'variables'. You would also
> not be concerned about reporting the possible duplicates multiple
> times.
>
> ALSO - you could warn the data entry staff, as the possible error is
> about to be commited, allowing for a much more efficient
> correction/followup.

I agree. If the OP had been more straight-forward in the first place it could have saved a lot of time and misdirection.

My feeling is that anything in a relational database that is unconstrained is, by definition, a memo field. If management doesn't like that they should go back to COBOL or RPGII.

-- 
Puget Sound Oracle Users Group