Numbering groups of rows?

create table t (colA int, colB char(1), colC int)
insert into t(colB) Values('C')
insert into t(colB) Values('C')
insert into t(colB) Values('C')
insert into t(colB) Values('C')
insert into t(colB) Values('A')
insert into t(colB) Values('A')
insert into t(colB) Values('A')
insert into t(colB) Values('A')
insert into t(colB) Values('B')
insert into t(colB) Values('B')
insert into t(colB) Values('B')
insert into t(colB) Values('B')
update t set colc =
(select count(*) from t as a where t.colb >= a.colb)
yields
NULL C 12
NULL C 12
NULL C 12
NULL C 12
NULL A 4
NULL A 4
NULL A 4
NULL A 4
NULL B 8
NULL B 8
NULL B 8
NULL B 8
how can I make it yield
NULL C 3
NULL C 3
NULL C 3
NULL C 3
NULL A 1
NULL A 1
NULL A 1
NULL A 1
NULL B 2
NULL B 2
NULL B 2
NULL B 2
Thanks,
Ronupdate t set colc =
(select count(DISTINCT a.colb) from t as a where t.colb >= a.colb)
Jacco Schalkwijk
SQL Server MVP
"Ron" <anonymous@.discussions.microsoft.com> wrote in message
news:01b401c50fb9$5a806ee0$a501280a@.phx.gbl...
> create table t (colA int, colB char(1), colC int)
> insert into t(colB) Values('C')
> insert into t(colB) Values('C')
> insert into t(colB) Values('C')
> insert into t(colB) Values('C')
> insert into t(colB) Values('A')
> insert into t(colB) Values('A')
> insert into t(colB) Values('A')
> insert into t(colB) Values('A')
> insert into t(colB) Values('B')
> insert into t(colB) Values('B')
> insert into t(colB) Values('B')
> insert into t(colB) Values('B')
> update t set colc =
> (select count(*) from t as a where t.colb >= a.colb)
> yields
> NULL C 12
> NULL C 12
> NULL C 12
> NULL C 12
> NULL A 4
> NULL A 4
> NULL A 4
> NULL A 4
> NULL B 8
> NULL B 8
> NULL B 8
> NULL B 8
> how can I make it yield
> NULL C 3
> NULL C 3
> NULL C 3
> NULL C 3
> NULL A 1
> NULL A 1
> NULL A 1
> NULL A 1
> NULL B 2
> NULL B 2
> NULL B 2
> NULL B 2
> Thanks,
> Ron
>|||On Thu, 10 Feb 2005 13:42:01 -0800, Ron wrote:

>update t set colc =
>(select count(*) from t as a where t.colb >= a.colb)
>yields
(snip)
>how can I make it yield
(snip)
Hi Ron,
Better not to store this information at all - you'll find yourself
constantly fighting to keep the rankingf column current after each
modification to the underlying data. It's better to drop the column from
the table and create a view to calculate it.
If you MUST do it in an update, try
update t set colc =
(select count(distinct a.colb) from t as a where t.colb >= a.colb)
Best, Hugo
--
(Remove _NO_ and _SPAM_ to get my e-mail address)|||Thanks very much. That worked well. But I also tried
using this in a select statement as follows for ranking
the values A, B, C
select a.colB, count(distinct a.colb) as colC
from t as a inner join t as b
on a.colB >= b.colB
group by a.colB
yielded this ranking
A 1
B 1
C 1
without the Distinct keyword I get this ranking
A 16
B 32
C 48
But I would like to get a ranking as follows
A 1
B 2
C 3
I ask this because I am trying to understand the sql logic
to achieve these results. Hopefully, after I do enough of
these kinds of queries I will get the idea how they work.
May I ask how I could achieve the ranking from result3?
Thanks again,
Ron

>--Original Message--
> update t set colc =
>(select count(DISTINCT a.colb) from t as a where t.colb
>= a.colb)
>
>--
>Jacco Schalkwijk
>SQL Server MVP
>
>"Ron" <anonymous@.discussions.microsoft.com> wrote in
message
>news:01b401c50fb9$5a806ee0$a501280a@.phx.gbl...
>
>.
>|||Ron,
Try count(distinct b.colb) instead of count(distinct a.colb). I suspect
that's what you had in mind.
Steve Kass
Drew University
Ron wrote:
>Thanks very much. That worked well. But I also tried
>using this in a select statement as follows for ranking
>the values A, B, C
>select a.colB, count(distinct a.colb) as colC
>from t as a inner join t as b
>on a.colB >= b.colB
>group by a.colB
>yielded this ranking
>A 1
>B 1
>C 1
>without the Distinct keyword I get this ranking
>A 16
>B 32
>C 48
>But I would like to get a ranking as follows
>A 1
>B 2
>C 3
>I ask this because I am trying to understand the sql logic
>to achieve these results. Hopefully, after I do enough of
>these kinds of queries I will get the idea how they work.
>May I ask how I could achieve the ranking from result3?
>Thanks again,
>Ron
>
>
>message
>|||On Thu, 10 Feb 2005 14:21:58 -0800, Ron wrote:

>Thanks very much. That worked well. But I also tried
>using this in a select statement as follows for ranking
>the values A, B, C
>select a.colB, count(distinct a.colb) as colC
>from t as a inner join t as b
>on a.colB >= b.colB
>group by a.colB
Hi Ron,
Try this one instead:
select a.colB, count(distinct b.colb) as colC
from t as a inner join t as b
on a.colB >= b.colB
group by a.colB
(Note: only one letter weas changed!!)
Best, Hugo
--
(Remove _NO_ and _SPAM_ to get my e-mail address)|||Thanks very much. I am also trying to figure out how to
rank A, B, C
select a.colB, count(distinct a.colb) as colC
from t as a inner join t as b
on a.colB >= b.colB
group by a.colB
yielded this ranking
A 1
B 1
C 1
without the Distinct keyword I get this ranking
A 16
B 32
C 48
But I would like to get a ranking as follows
A 1
B 2
C 3
May I ask how I could achieve the ranking from result3?
This way, as you say, I don't really store the ranks, just
retrieve them dynamically.
Thanks again,
Ron

>--Original Message--
>On Thu, 10 Feb 2005 13:42:01 -0800, Ron wrote:
>
>(snip)
>(snip)
>Hi Ron,
>Better not to store this information at all - you'll find
yourself
>constantly fighting to keep the rankingf column current
after each
>modification to the underlying data. It's better to drop
the column from
>the table and create a view to calculate it.
>If you MUST do it in an update, try
>update t set colc =
>(select count(distinct a.colb) from t as a where t.colb
>= a.colb)
>Best, Hugo
>--
>(Remove _NO_ and _SPAM_ to get my e-mail address)
>.
>|||Thanks all. I am sort of starting to get the idea. But I
wonder if I could belabor this thing one more notch:
Instead of using distinct is it possible to plant a group
by query in there? Pseudocode here:
select a.colB, count(select a.colb from a group by a.colb)
as colC from t as a inner join t as b
on a.colB >= b.colB group by a.colB
Again, I just ask because I don't really know all the
rules for t-sql, let alone the tricks. I am guessing that
t-sql does not allow Selects inside of Count(..)
Thanks again,
Ron

>--Original Message--
>Thanks very much. That worked well. But I also tried
>using this in a select statement as follows for ranking
>the values A, B, C
>select a.colB, count(distinct a.colb) as colC
>from t as a inner join t as b
>on a.colB >= b.colB
>group by a.colB
>yielded this ranking
>A 1
>B 1
>C 1
>without the Distinct keyword I get this ranking
>A 16
>B 32
>C 48
>But I would like to get a ranking as follows
>A 1
>B 2
>C 3
>I ask this because I am trying to understand the sql
logic
>to achieve these results. Hopefully, after I do enough
of
>these kinds of queries I will get the idea how they
work.
>May I ask how I could achieve the ranking from result3?
>Thanks again,
>Ron
>
>message
>.
>|||On Thu, 10 Feb 2005 14:45:11 -0800, Ron wrote:

>Thanks all. I am sort of starting to get the idea. But I
>wonder if I could belabor this thing one more notch:
>Instead of using distinct is it possible to plant a group
>by query in there? Pseudocode here:
>select a.colB, count(select a.colb from a group by a.colb)
>as colC from t as a inner join t as b
>on a.colB >= b.colB group by a.colB
Hi Ron,
This code won't work. I'm sure there is some way to do this with a group
by in the subquery, but it's not trivial and it'll be more complex than
the version with DISTINCT that I suggested.
Best, Hugo
--
(Remove _NO_ and _SPAM_ to get my e-mail address)|||>> Instead of using distinct is it possible to plant a group by query
in there? <<
This will use a GROUP BY and get you a bit more information in the VIEW
aggregate functions. I am not sure that there is any advantage. .
CREATE TABLE Foobar (letter CHAR(1) NOT NULL);
INSERT INTO Foobar (letter) VALUES('C');
INSERT INTO Foobar (letter) VALUES('C');
INSERT INTO Foobar (letter) VALUES('C');
INSERT INTO Foobar (letter) VALUES('C');
INSERT INTO Foobar (letter) VALUES('A');
INSERT INTO Foobar (letter) VALUES('A');
INSERT INTO Foobar (letter) VALUES('A');
INSERT INTO Foobar (letter) VALUES('A');
INSERT INTO Foobar (letter) VALUES('B');
INSERT INTO Foobar (letter) VALUES('B');
INSERT INTO Foobar (letter) VALUES('B');
INSERT INTO Foobar (letter) VALUES('B');
CREATE VIEW FoobarReport (letter, occurs, place)
AS
SELECT F1.letter, COUNT(*),
(SELECT COUNT (DISTINCT F2.letter)
FROM Foobar AS F2
WHERE F2.letter <= F1.letter)
FROM Foobar AS F1
GROUP BY F1.letter;