You don't mention your RDBMS.
But some RDBMS (and I presume all internally) include a unique row
identifier which you can access (e.g Ingres - tid )
If this is available, then the rows are no longer 'unique'
and you can decide which row you want to get rid of
e.g max/min(tid) group by having c
I did post previously, but assume no-one out there is sad enough to still be using
ingres with something as dynamic as perl.
My make test fails, either with an undefined symbol
mutex_init
if -lthread is specified, or undefined symbol
finite
when -Kthread is used.
As below
ins