Re: [SQL] DISTINCT count(*), possible?

On Sun, 20 Jun 1999, Herouth Maoz wrote:

> > Out of curiosity, more than anything else.
> > Would the group by solve this problem?
>
> No. What you need is the number of groups, not the number of members in
> each group, which is what group by provides.
>
> Herouth

Good morning Herouth,

I missed the earlier post(s) on this thread (scrambling to meet
a 01 July go-live date) but the subj caught my eye today as just
yesterday I tried to force 6.3.2 to do a select count(distinct id)
and was rebuffed!

Does this work in 6.5.x?

Also - is there a pg way to do an oracle nvl() type assignment of a
char str to a NULL value during a query (e.g., 'unmarried');

Finally, I have a annoyance. My predecessor apparently was unaware or
distrustful of the text data type. Instead he did this:

note_id int4
line_no int4
noteline char(65)

The user enters lines of text, they are numbered and assigned the (same)
note_id which links them to one note. Doing a select they are reassembled
into one note. (I prefer the text data type as I'm not fond of extra
typing... ;-)

Now, I am porting this old PROGRESS db to Postgres and I need to concat
these char strs into a text attribute. First I dumped and reloaded,
changing the attr holding each str from char() to text. Now I am clumsily
trying to use the concat operator to select into the new table.

INSERT INTO assessment (assets, debits)
SELECT asset1 || asset2 AS assets, debit1 || debit2 AS debits
FROM oldtab
WHERE id = xyz AND date = '04-01-1999';

Sometimes it works and mostly it doesn't.
Would you know of a better approach?

Many thanks,
Tom

------- North Richmond Community Mental Health Center -------

Thomas Good MIS Coordinator, Senior DBA
Vital Signs: tomg@ { admin | q8 } .nrnet.org
Phone: 718-354-5528
Fax: 718-354-5056

/* Member: Computer Professionals For Social Responsibility */