I have a “logging” table. It records the user, login time, and other
things.

I want to create table within a statement that returns one row for
each user. This row is the row containing the user’s oldest date

Select blah blah from table x, (select user, date, other, info from
logging where date is max) oldrows
where x.date < oldrows.date
and x.user = oldrows.user

It’s building the table here that looks very difficult. This would be
easy enough to do in pl/sql, but I rather do it this way, if it can
even be done. This kind of thing is going to come up a lot.

Thank you

On Sep 16, 6:37*pm, okey <oldyor... (AT) yahoo (DOT) com> wrote:
I thought this might be a way, another subselect to find max date.
As in:

select blah, blah from logtable t, (select distinct user, max(date)
group by user) mx where mx.user = t.user ...

I'd have to assume the distinct user date was indeed unique. Not
there yet unless I create yet another subselect to select that max
date record and group it with a max unique key - any one would do.

You'd think there was something better.

On Sep 16, 8:00*pm, okey <oldyor... (AT) yahoo (DOT) com> wrote:
Who can say siince you haven't posted a create table statement and
some sample data. We have no idea what your data looks like so how
can we suggest an alternative?


David Fitzjarrell

Op 17-9-2010 1:37, okey schreef:

select user, min(login_time)
from x
group by user

?

Shakespeare