Yeah, I could. And you are right, *never* in an application...but the minute I write something definitive like "never use a unique IDX in an application" a user will call me and I'll need to use one :)


>Could you just SELECT DISTINCT ... ? Basically, there is nothing against using these indexes for ad-hoc operations, but seeing them in application, it's not good
>
>>Josh ---
>>
>>You are correct, however I have found Unique indexes to be invaluable when developing. A lot of my work involves redesigning databases and migrating legacy tables into VFP. When I go to normalize a column, I create a temporary compact unique IDX to see the range of values I have to be concerned with ... I just couldn't do that with a Candidate key (even if I wanted to create a CDX)
>>
>>>
>>>That's true, but a unique index is not the same as a candidate index. It doesn't keep you from adding duplicate values, it just ensures that only one record will appear in the index for each distinct key value.
>>>
>>>>Rock ---
>>>>
>>>>Yes, it's still needed. If you are creating a standalone index (IDX), Candidate is not an option.
>>>>