Re: delete duplicates takes too long

On Sat, Apr 25, 2009 at 2:51 AM, Alvaro Herrera
 wrote:
> In that case you can create a partial unique index:
>
> create index foo on cdr_ama_stat (abonado_a, abonado_b, fecha_llamada,
> duracion) where processed = 2;

Of course, the unique is missing in this one:

CREATE UNIQUE INDEX foo ON cdr_ama_stat (abonado_a, abonado_b, fecha_llamada,
duracion) WHERE processed = 2;