> I have 96,000 records in a postgres table that has an unpopulated field
> for
> datetime. I need the value in that field to be unique, but the ACTUAL
> value isn't really that important. I'd like to begin with the first
> record, enter a date and time, and then increment that value by one second
>
> for the next field, one second for the next, and so on. Can I do this
> directly from within psql?? I could do this in dbase w/ no problem, but
> I'm really not that familiar with sql syntax. . .
>
> jt
>
You might be able to do it with a creative use of a sequence.
create sequence my_seq;
create table test ( my_d datetime PRIMARY KEY DEFAULT
now()+(nextval('my_seq') || ' seconds'), ...)
You still have problems with duplicates happening as in my reply to your
othe message.
you can take the default (now()+(nextval('my_seq') || ' seconds')) and use
it in an INSERT statement to the same effect.
-DEJ