How can I combine 2 SELECT statements into 1 ?

Please consider the following SELECT statements; is it possible to combine
them into a sole SELECT statement returning 4 columns ? Thank you.
SELECT YEAR(DhEnregistrementIntervention) AS ItvNumAnnee
, MONTH(DhEnregistrementIntervention) AS ItvNumMois
, COUNT(NumIntervention) AS ItvQteDemande
FROM dbo.InterventionDI
GROUP BY YEAR(DhEnregistrementIntervention),
MONTH(DhEnregistrementIntervention)
ORDER BY YEAR(DhEnregistrementIntervention),
MONTH(DhEnregistrementIntervention);
SELECT YEAR(DhEnregistrementIntervention) AS ItvAnneeEnregistrement
, MONTH(DhEnregistrementIntervention) AS ItvMoisEnregistrement
, ROUND(AVG(DATEDIFF(DAY, DhEnregistrementIntervention,
DhClotureIntervention)), 0) AS ItvDuree
FROM dbo.InterventionDI
WHERE (CodEtatIntervention = 'F')
GROUP BY YEAR(DhEnregistrementIntervention),
MONTH(DhEnregistrementIntervention)
ORDER BY YEAR(DhEnregistrementIntervention),
MONTH(DhEnregistrementIntervention);
--
Gilbert.HI
Use UNION or UNION ALL
Cheers
vinu
"Gilbert" <gilbert@.nospam.nospam> wrote in message
news:3B7C228D-3913-49EF-85E3-776520901701@.microsoft.com...
> Please consider the following SELECT statements; is it possible to combine
> them into a sole SELECT statement returning 4 columns ? Thank you.
> SELECT YEAR(DhEnregistrementIntervention) AS ItvNumAnnee
> , MONTH(DhEnregistrementIntervention) AS ItvNumMois
> , COUNT(NumIntervention) AS ItvQteDemande
> FROM dbo.InterventionDI
> GROUP BY YEAR(DhEnregistrementIntervention),
> MONTH(DhEnregistrementIntervention)
> ORDER BY YEAR(DhEnregistrementIntervention),
> MONTH(DhEnregistrementIntervention);
> SELECT YEAR(DhEnregistrementIntervention) AS ItvAnneeEnregistrement
> , MONTH(DhEnregistrementIntervention) AS ItvMoisEnregistrement
> , ROUND(AVG(DATEDIFF(DAY, DhEnregistrementIntervention,
> DhClotureIntervention)), 0) AS ItvDuree
> FROM dbo.InterventionDI
> WHERE (CodEtatIntervention = 'F')
> GROUP BY YEAR(DhEnregistrementIntervention),
> MONTH(DhEnregistrementIntervention)
> ORDER BY YEAR(DhEnregistrementIntervention),
> MONTH(DhEnregistrementIntervention);
> --
> Gilbert.|||Sorry but, as far as I know, with UNION I will get more lines with always 3
columns. This is not why I need. I need the same number of lines
(approximatively) with 4 columns.
--
Gilbert.
"vinu" wrote:

> HI
> Use UNION or UNION ALL
> Cheers
> vinu
>
> "Gilbert" <gilbert@.nospam.nospam> wrote in message
> news:3B7C228D-3913-49EF-85E3-776520901701@.microsoft.com...
>
>|||Please post DDL, sample data and expected results.
You might also consider joining the two queries as derived tables on common
columns (e.g. ItvNumAnee and ItvNumMois), but don't let this suggestion fool
you into thinking I have any idea what you really want. Stop us guessing and
explain your goals.
ML
http://milambda.blogspot.com/|||I thought that giving the SELECT statements was enough details; sorry. I
intend to do monthly statistics on a table storing user requests. The first
query gives the number of requests per month. The second query gives the
average delay of the answer per month, for all requests that are closed (thi
s
is why there is a WHERE clause). If I execute 2 independent SELECT statement
s
I receive 2 tables with 3 columns, each line displaying the value for a
specific month. What I would like to do is to execute only 1 SELECT, giving
me 1 table of 4 columns (year, month, number of requests, and average delay)
,
each line displaying the values for a specific month. Thank you for your hel
p.
--
Gilbert.
"ML" wrote:

> Please post DDL, sample data and expected results.
> You might also consider joining the two queries as derived tables on commo
n
> columns (e.g. ItvNumAnee and ItvNumMois), but don't let this suggestion fo
ol
> you into thinking I have any idea what you really want. Stop us guessing a
nd
> explain your goals.
>
> ML
> --
> http://milambda.blogspot.com/|||Ok, I see. But we still need to understand how the data in the two queries
can be joined to form one valid result-set.
That's why we need to see the DDL and sample data:
http://www.aspfaq.com/etiquette.asp?id=5006
ML
http://milambda.blogspot.com/|||OK. Here is the DDL (simplified by hand to remove unuseful columns and
constraints) :
CREATE TABLE [InterventionDI] (
[NumIntervention] [int] IDENTITY (1, 1) NOT NULL ,
[DhEnregistrementIntervention] [datetime] NOT NULL CONSTRAINT
[DF_InterventionDI_DhEnregistrementInter
vention] DEFAULT (getdate()),
[DhClotureIntervention] [datetime] NULL ,
[CodEtatIntervention] [char] (1) COLLATE SQL_Latin1_General_CP1_CI_AI NOT
NULL,
CONSTRAINT [PK_InterventionDI] PRIMARY KEY NONCLUSTERED
(
[NumIntervention]
) WITH FILLFACTOR = 90 ON [PRIMARY] ,
) ON [PRIMARY]
GO
Here is a set on 10 records (INSERT created by hand) :
INSERT INTO dbo.InterventionDI (NumIntervention,
DhEnregistrementIntervention, DhClotureIntervention, CodEtatIntervention)
VALUES (18, '13/09/2002 09:52:18', '20/09/2002 16:53:57', 'F')
INSERT INTO dbo.InterventionDI (NumIntervention,
DhEnregistrementIntervention, DhClotureIntervention, CodEtatIntervention)
VALUES (19, '13/09/2002 09:53:18', '21/09/2002 16:53:57', 'F')
INSERT INTO dbo.InterventionDI (NumIntervention,
DhEnregistrementIntervention, DhClotureIntervention, CodEtatIntervention)
VALUES (20, '14/09/2002 09:54:18', '20/09/2002 16:53:57', 'F')
INSERT INTO dbo.InterventionDI (NumIntervention,
DhEnregistrementIntervention, DhClotureIntervention, CodEtatIntervention)
VALUES (21, '14/09/2002 09:55:18', '15/09/2002 16:53:57', 'O')
INSERT INTO dbo.InterventionDI (NumIntervention,
DhEnregistrementIntervention, DhClotureIntervention, CodEtatIntervention)
VALUES (22, '14/09/2002 09:56:18', '15/09/2002 16:53:57', 'F')
INSERT INTO dbo.InterventionDI (NumIntervention,
DhEnregistrementIntervention, DhClotureIntervention, CodEtatIntervention)
VALUES (23, '14/09/2002 09:57:18', '18/09/2002 16:53:57', 'F')
INSERT INTO dbo.InterventionDI (NumIntervention,
DhEnregistrementIntervention, DhClotureIntervention, CodEtatIntervention)
VALUES (24, '13/10/2002 09:58:18', '20/09/2003 16:53:57', 'F')
INSERT INTO dbo.InterventionDI (NumIntervention,
DhEnregistrementIntervention, DhClotureIntervention, CodEtatIntervention)
VALUES (25, '13/10/2002 09:59:18', '20/09/2004 16:53:57', 'F')
INSERT INTO dbo.InterventionDI (NumIntervention,
DhEnregistrementIntervention, DhClotureIntervention, CodEtatIntervention)
VALUES (26, '13/01/2003 09:12:18', '20/09/2005 16:53:57', 'F')
INSERT INTO dbo.InterventionDI (NumIntervention,
DhEnregistrementIntervention, DhClotureIntervention, CodEtatIntervention)
VALUES (27, '13/01/2003 09:22:18', '20/09/2006 16:53:57', 'F')
Thank you.
--
Gilbert.
"ML" wrote:

> Ok, I see. But we still need to understand how the data in the two queries
> can be joined to form one valid result-set.
> That's why we need to see the DDL and sample data:
> http://www.aspfaq.com/etiquette.asp?id=5006
>
> ML
> --
> http://milambda.blogspot.com/|||Based on what the two queries have in common this might give the correct
result:
select Demande.Anee
,Demande.Mois
,Demande.ItvQteDemande
,Duree.ItvDuree
from (
SELECT YEAR(DhEnregistrementIntervention) AS ItvNumAnnee
, MONTH(DhEnregistrementIntervention) AS ItvNumMois
, COUNT(NumIntervention) AS ItvQteDemande
FROM dbo.InterventionDI
GROUP BY YEAR(DhEnregistrementIntervention),
MONTH(DhEnregistrementIntervention)
) Demande (Anee, Mois, ItvQteDemande)
inner join (
SELECT YEAR(DhEnregistrementIntervention) AS ItvAnneeEnregistrement
, MONTH(DhEnregistrementIntervention) AS ItvMoisEnregistrement
, ROUND(AVG(DATEDIFF(DAY, DhEnregistrementIntervention,
DhClotureIntervention)), 0) AS ItvDuree
FROM dbo.InterventionDI
WHERE (CodEtatIntervention = 'F')
GROUP BY YEAR(DhEnregistrementIntervention),
MONTH(DhEnregistrementIntervention)
) Duree (Anee, Mois, ItvDuree)
on (Duree.Anee = Demande.Anee)
and (Duree.Mois = Demande.Mois)
order by Demande.Anee
,Demande.Mois
Please, validate this against a sufficient amount of cases before using it
in production.
ML
http://milambda.blogspot.com/|||UNTESTED
SELECT YEAR(DhEnregistrementIntervention) AS ItvNumAnnee
, MONTH(DhEnregistrementIntervention) AS ItvNumMois
, COUNT(NumIntervention) AS ItvQteDemande
, (
case when CodEtatIntervention = 'F' then
ROUND(AVG(DATEDIFF(DAY, DhEnregistrementIntervention,
DhClotureIntervention)), 0)
ELSE NULL END
)AS ItvDuree
FROM dbo.InterventionDI
GROUP BY YEAR(DhEnregistrementIntervention),
MONTH(DhEnregistrementIntervention)
ORDER BY YEAR(DhEnregistrementIntervention),
MONTH(DhEnregistrementIntervention);
"Gilbert" <gilbert@.nospam.nospam> wrote in message
news:3B7C228D-3913-49EF-85E3-776520901701@.microsoft.com...
> Please consider the following SELECT statements; is it possible to combine
> them into a sole SELECT statement returning 4 columns ? Thank you.
> SELECT YEAR(DhEnregistrementIntervention) AS ItvNumAnnee
> , MONTH(DhEnregistrementIntervention) AS ItvNumMois
> , COUNT(NumIntervention) AS ItvQteDemande
> FROM dbo.InterventionDI
> GROUP BY YEAR(DhEnregistrementIntervention),
> MONTH(DhEnregistrementIntervention)
> ORDER BY YEAR(DhEnregistrementIntervention),
> MONTH(DhEnregistrementIntervention);
> SELECT YEAR(DhEnregistrementIntervention) AS ItvAnneeEnregistrement
> , MONTH(DhEnregistrementIntervention) AS ItvMoisEnregistrement
> , ROUND(AVG(DATEDIFF(DAY, DhEnregistrementIntervention,
> DhClotureIntervention)), 0) AS ItvDuree
> FROM dbo.InterventionDI
> WHERE (CodEtatIntervention = 'F')
> GROUP BY YEAR(DhEnregistrementIntervention),
> MONTH(DhEnregistrementIntervention)
> ORDER BY YEAR(DhEnregistrementIntervention),
> MONTH(DhEnregistrementIntervention);
> --
> Gilbert.|||Wonderful ! Thank you for having teached me something new.
--
Gilbert.
"ML" wrote:

> Based on what the two queries have in common this might give the correct
> result:
> select Demande.Anee
> ,Demande.Mois
> ,Demande.ItvQteDemande
> ,Duree.ItvDuree
> from (
> SELECT YEAR(DhEnregistrementIntervention) AS ItvNumAnnee
> , MONTH(DhEnregistrementIntervention) AS ItvNumMois
> , COUNT(NumIntervention) AS ItvQteDemande
> FROM dbo.InterventionDI
> GROUP BY YEAR(DhEnregistrementIntervention),
> MONTH(DhEnregistrementIntervention)
> ) Demande (Anee, Mois, ItvQteDemande)
> inner join (
> SELECT YEAR(DhEnregistrementIntervention) AS ItvAnneeEnregistrement
> , MONTH(DhEnregistrementIntervention) AS ItvMoisEnregistrement
> , ROUND(AVG(DATEDIFF(DAY, DhEnregistrementIntervention,
> DhClotureIntervention)), 0) AS ItvDuree
> FROM dbo.InterventionDI
> WHERE (CodEtatIntervention = 'F')
> GROUP BY YEAR(DhEnregistrementIntervention),
> MONTH(DhEnregistrementIntervention)
> ) Duree (Anee, Mois, ItvDuree)
> on (Duree.Anee = Demande.Anee)
> and (Duree.Mois = Demande.Mois)
> order by Demande.Anee
> ,Demande.Mois
> Please, validate this against a sufficient amount of cases before using it
> in production.
>
> ML
> --
> http://milambda.blogspot.com/sql

How best to do remote site replication?

It has been suggested that we consider how to replicate our production
server to a remote site that does not have a dedicated connection to our
production environment. Any and all suggestions are appreciated.
TIA
Michael MacGregor
Database Architect
SalesDriver
Michael,
assuming there's no direct connection, this type of setup is usually done
through a VPN.
Here is a reference to help you:
http://support.microsoft.com/?id=321822.
Regards,
Paul Ibison
|||Additionally, I should mention that pull subscriptions would be most
suitable.
Regards,
Paul Ibison
|||Thanks Paul.
Totally new to replication and our IT guy is proposing some really weird and
wacky ideas on how to do this, one of which is low-level bit comparison, and
subsequent updating, of backup files to use to restore a database. Makes me
shake my head in wonder.
Michael MacGregor
Database Architect
SalesDriver
|||Michael,
if this is for disaster recovery you might want to consider log-shipping.
There have been a few threads recently (last 2 weeks) comparing this to
transactional replication which might be helpful, and it is a standard way
of maintaining a database copy. Using propriety solutions like
bit-comparisons would no doubt not be supported by ms, and presumably would
require a comparison of the full database backup. Also, log shipping will
just require that you transmit the changes which are in the log to the
remote site rather than full backups.
HTH,
Paul Ibison
|||To be honest, I'm not too sure what problem this is supposed to address. We
have an active/passive SQL Server cluster which has only been required once
in over three years. In addition there is a fairly standard backup/recovery
plan in effect with regular backups, both full and transactional, performed.
The IT guy has got the idea of having another SQL Server installed at a
remote site, to act as what I'm not sure, but wants to have it as close to
the production server as possible.
I wondered about log-shipping as well, but I'm not even sure that this idea
is going to fly anyway, but thought I'd do a little research of my own
before this issue gets raised on a more formal basis, at which point in time
I can address it better then.
Would you be able to provide me with the subject titles of the threads you
refer to?
Michael MacGregor
|||Michael,
the main thread was entitled "Disaster recovery - Use Replication or Log
shipping".
HTH,
Paul Ibison