>>< SNIP >
>>Nadya,
>>
>>I think you should write FAQ 'How to delete duplicate records'. Include all four ideas: Message #
565099, Message #
565029, Message #
565097 and Message #
565101 with explanation of there Pros and Cons.
>
>Ok, I'll do. BTW, another method would be:
>
>select distinct * from originaltable into myNewTable
>delete originalTable
>rename file NewTable to OriginalTable, if we're talking about free tables
>or zap original and append.
It'll only work for exact duplicates (all fields). Using GROUP BY instead of DISTINCT can fix it.
>However, I believe, the algorithm, I finally described with your corrections is the optimum here.
You'll have to do some bench marking to prove it :)
I still think that all of them can be useful.
--sb--