My point exactly. That's why I adviced checking for duplicates _after_ removing indices and primary keys on problematic tables. I've never tried the method proposed by Svein, but if it really omits indices, it's better than mine (more subtle and doesn't mess with metadata).
regards Tomasz On 2011-11-17 08:10, Svein Erling Tysvær wrote: >> Hi Tomasz. I already checked the source database for duplicated entries, >> and found none, but i did not tried to verify the database with gfix. > > Exactly how did you check it for duplicate entries? The point is that if > there's a problem with an index or key, then doing > > SELECT MyIndexedField, count(*) > FROM MyTable > GROUP BY 1 > HAVING count(*)> 1 > > will use the index/key and not discover the duplicate entry, whereas > > SELECT MyIndexedField+0, count(*) > FROM MyTable > GROUP BY 1 > HAVING count(*)> 1 > > cannot use the index and will find such duplicates. > > HTH, > Set > -- __--==============================--__ __--== Tomasz Tyrakowski ==--__ __--== SOL-SYSTEM ==--__ __--== http://www.sol-system.pl ==--__ __--==============================--__
