November 15, 2012 at 5:52 pm
--I believe the DDL and my commented out notes and questions should explain what I am hoping to achieve
CREATE TABLE #Companies (CompanyID int)
INSERT INTO #Companies (CompanyID) VALUES (1)
INSERT INTO #Companies (CompanyID) VALUES (2)
INSERT INTO #Companies (CompanyID) VALUES (3)
CREATE TABLE #CompanyGroups (CompanyGroupID char(1), CompanyID int)
INSERT INTO #CompanyGroups (CompanyGroupID,CompanyID) VALUES ('A',1)
INSERT INTO #CompanyGroups (CompanyGroupID,CompanyID) VALUES ('A',2)
DECLARE @CompanyID int
DECLARE @CompanyGroupID char(1)
--Scenario 1
--@CompanyID is populated, CompanyGroupID is NULL
SET @CompanyID = 1
SET @CompanyGroupID = NULL
IF @CompanyID IS NOT NULL
BEGIN
SELECT * FROM #Companies WHERE CompanyID = @CompanyID
END
--Scenario 2
--@CompanyID is NULL, CompanyGroupID is populated
SET @CompanyID = NULL
SET @CompanyGroupID = 'A'
IF @CompanyID IS NULL
BEGIN
SELECT * FROM #Companies WHERE CompanyID IN(Select CompanyID FROM #CompanyGroups WHERE CompanyGroupID = @CompanyGroupID)
END
--So then I can do if else
IF @CompanyID IS NOT NULL
BEGIN
SELECT * FROM #Companies WHERE CompanyID = @CompanyID
END
ELSE
BEGIN
SELECT * FROM #Companies WHERE CompanyID IN(Select CompanyID FROM #CompanyGroups WHERE CompanyGroupID = @CompanyGroupID)
END
--SELECT * FROM #Companies
--SELECT * FROM #CompanyGroups
DROP TABLE #Companies
DROP TABLE #CompanyGroups
--The question: My select statement is complicated and I'd prefer not to duplicate it in each part of the if/else statement.
--Can I build a single SELECT statement that will determine if @CompanyID IS NOT NULL and select the appropriate records??
November 15, 2012 at 8:22 pm
It's a bit ugly ... but you could do something like
;with companyNotNull as (
SELECT * FROM #Companies WHERE CompanyID = @CompanyID
),
companyNull as (
Select CompanyID FROM #CompanyGroups WHERE CompanyGroupID = @CompanyGroupID
)
SELECT *
FROM companyNotNull
WHERE @companyID is not null
UNION ALL
SELECT *
FROM companyNull
WHERE @companyID is null
Of course this replicates the if ... then ... else structure. If both values are set then the companyID takes precedence.
November 15, 2012 at 8:41 pm
In addition to MickyT's suggestion, you could do a catch all query or the dynamic SQL method suggested in this article by Gail Shaw:
http://sqlinthewild.co.za/index.php/2009/03/19/catch-all-queries/
Watch out for SQL injection if the parameter values come from the UI if you use the dynamic SQL method though.
My thought question: Have you ever been told that your query runs too fast?
My advice:
INDEXing a poor-performing query is like putting sugar on cat food. Yeah, it probably tastes better but are you sure you want to eat it?
The path of least resistance can be a slippery slope. Take care that fixing your fixes of fixes doesn't snowball and end up costing you more than fixing the root cause would have in the first place.
Need to UNPIVOT? Why not CROSS APPLY VALUES instead?[/url]
Since random numbers are too important to be left to chance, let's generate some![/url]
Learn to understand recursive CTEs by example.[/url]
[url url=http://www.sqlservercentral.com/articles/St
November 15, 2012 at 10:14 pm
PLS TRY BELOW CODE
CREATE TABLE #Companies (CompanyID int)
INSERT INTO #Companies (CompanyID) VALUES (1)
INSERT INTO #Companies (CompanyID) VALUES (2)
INSERT INTO #Companies (CompanyID) VALUES (3)
CREATE TABLE #CompanyGroups (CompanyGroupID char(1), CompanyID int)
INSERT INTO #CompanyGroups (CompanyGroupID,CompanyID) VALUES ('A',1)
INSERT INTO #CompanyGroups (CompanyGroupID,CompanyID) VALUES ('A',2)
DECLARE @CompanyID int
DECLARE @CompanyGroupID char(1)
--SET @CompanyID = 1
--SET @CompanyGroupID = NULL
SET @CompanyID = NULL
SET @CompanyGroupID = 'A'
SELECT * FROM #Companies WHERE CompanyID = (CASE WHEN @CompanyID IS NULL THEN CompanyID ELSE @CompanyID END)
AND CompanyID IN(Select CompanyID FROM #CompanyGroups WHERE CompanyGroupID = CASE WHEN @CompanyGroupID IS NULL THEN CompanyGroupID ELSE @CompanyGroupID END)
November 15, 2012 at 10:22 pm
mickyT (11/15/2012)
It's a bit ugly ... but you could do something like
;with companyNotNull as (SELECT * FROM #Companies WHERE CompanyID = @CompanyID
),
companyNull as (
Select CompanyID FROM #CompanyGroups WHERE CompanyGroupID = @CompanyGroupID
)
SELECT *
FROM companyNotNull
WHERE @companyID is not null
UNION ALL
SELECT *
FROM companyNull
WHERE @companyID is null
Of course this replicates the if ... then ... else structure. If both values are set then the companyID takes precedence.
It occurred to me on lookback that this is going to fail if #Companies has more than one column or that column has a data type that is incompatible with CompanyID from #CompanyGroups.
My thought question: Have you ever been told that your query runs too fast?
My advice:
INDEXing a poor-performing query is like putting sugar on cat food. Yeah, it probably tastes better but are you sure you want to eat it?
The path of least resistance can be a slippery slope. Take care that fixing your fixes of fixes doesn't snowball and end up costing you more than fixing the root cause would have in the first place.
Need to UNPIVOT? Why not CROSS APPLY VALUES instead?[/url]
Since random numbers are too important to be left to chance, let's generate some![/url]
Learn to understand recursive CTEs by example.[/url]
[url url=http://www.sqlservercentral.com/articles/St
November 15, 2012 at 10:42 pm
I think this is the "normal" catch all query form for this.
SELECT *
FROM #Companies
WHERE (@CompanyID IS NULL OR CompanyID = @CompanyID) AND
(@CompanyGroupID IS NULL OR
CompanyID IN(Select CompanyID FROM #CompanyGroups WHERE CompanyGroupID = @CompanyGroupID))
My thought question: Have you ever been told that your query runs too fast?
My advice:
INDEXing a poor-performing query is like putting sugar on cat food. Yeah, it probably tastes better but are you sure you want to eat it?
The path of least resistance can be a slippery slope. Take care that fixing your fixes of fixes doesn't snowball and end up costing you more than fixing the root cause would have in the first place.
Need to UNPIVOT? Why not CROSS APPLY VALUES instead?[/url]
Since random numbers are too important to be left to chance, let's generate some![/url]
Learn to understand recursive CTEs by example.[/url]
[url url=http://www.sqlservercentral.com/articles/St
November 15, 2012 at 11:35 pm
Additional questions. What results should be for this?
CREATE TABLE #Companies (CompanyID int)
INSERT INTO #Companies (CompanyID) VALUES (1)
INSERT INTO #Companies (CompanyID) VALUES (2)
INSERT INTO #Companies (CompanyID) VALUES (3)
CREATE TABLE #CompanyGroups (CompanyGroupID char(1), CompanyID int)
INSERT INTO #CompanyGroups (CompanyGroupID,CompanyID) VALUES ('A',1)
INSERT INTO #CompanyGroups (CompanyGroupID,CompanyID) VALUES ('A',2)
a.
SET @CompanyID = 1
SET @CompanyGroupID = 'A'
b.
SET @CompanyID = 1
SET @CompanyGroupID = 'B'
c.
SET @CompanyID = 3
SET @CompanyGroupID = 'A'
d.
SET @CompanyID = 3
SET @CompanyGroupID = 'E'
e.
SET @CompanyID = NULL
SET @CompanyGroupID = NULL
:unsure:
November 15, 2012 at 11:36 pm
You could think about something like this:
SELECT Comp.*
FROM #Companies AS Comp
LEFT JOIN #CompanyGroups AS Gr ON Comp.CompanyID = Gr.CompanyID
WHERE Comp.CompanyID = @CompanyID
OR GR.CompanyGroupID = @CompanyGroupID
November 16, 2012 at 6:40 am
SELECT *
FROM
#Companies
WHERE CompanyID = @CompanyID
OR
(
CompanyID IS NULL AND
CompanyID IN(Select CompanyID FROM #CompanyGroups WHERE CompanyGroupID = @CompanyGroupID)
)
November 16, 2012 at 11:10 am
>>Additional questions. What results should be for this?
The parameters will be validated through an user interface so I don't expect any of those conditions to exist.
November 16, 2012 at 11:12 am
Thanks all. I went with dwain.c's "normal" catch all query form.
November 16, 2012 at 1:02 pm
dwain.c (11/15/2012)
It occurred to me on lookback that this is going to fail if #Companies has more than one column or that column has a data type that is incompatible with CompanyID from #CompanyGroups.
Oops ... that's what I get for rushing things probably should have done the following, which was what I had in my head at the time;with companyNotNull as (
SELECT * FROM #Companies WHERE CompanyID = @CompanyID
),
companyNull as (
Select CompanyID FROM #CompanyGroups WHERE CompanyGroupID = @CompanyGroupID
)
SELECT *
FROM companyNotNull
WHERE @companyID is not null
UNION ALL
SELECT *
FROM #Companies c
WHERE exists (SELECT 1 FROM companyNull n WHERE n.CompanyID = c.CompanyID)
November 18, 2012 at 5:35 pm
Post deleted. Coffee has not yet arrived.
My thought question: Have you ever been told that your query runs too fast?
My advice:
INDEXing a poor-performing query is like putting sugar on cat food. Yeah, it probably tastes better but are you sure you want to eat it?
The path of least resistance can be a slippery slope. Take care that fixing your fixes of fixes doesn't snowball and end up costing you more than fixing the root cause would have in the first place.
Need to UNPIVOT? Why not CROSS APPLY VALUES instead?[/url]
Since random numbers are too important to be left to chance, let's generate some![/url]
Learn to understand recursive CTEs by example.[/url]
[url url=http://www.sqlservercentral.com/articles/St
November 19, 2012 at 11:43 am
CELKO (11/19/2012)
... Companies are identified by a DUNS (we even get radio ads for DUNS here in Austin!);...
That is real delight! How cool it is! There is only one small problem here...
Are you sure that most of people on this planet have any idea about where Austin is?
I honestly believe, if everyone would knew that there is a radio ads for DUNS in your place, all of people would try to get to Austin to hear it straight away (may be youtube link at least)!
I can also believe that you have checked every single company in Austin and made sure they have their DUNS done. Including every independent pizza or burger shop...
Sorry to upset you, but unfortunately the rest of the world is not quite in Austin yet, therefore there are still some companies around (and people) who never heard about such a beautiful thing as DUNS, which allows to uniquely identify company on our small planet. Actually, I've worked for quite few large companies who do deal with a lot of corporate clients of any size, and I never, really NEVER! have seen any database where DUNS is used as PK. Actually, before you have mentioned it couple month ago, I have never even heard this word at all. But, of course, it is poor me, living on the outskirts of the world in the middle-of-no-where called London-village...
Oh lucky you...
I will now sleep and dream about real paradise on the earth called Austin, where even radio informs happy citizens about DUNS...
:hehe:
By the way, for many of us rows and records are the same... I'm so sorry... 😉
And to OP, just stick with INT CompanyId's PKs, it's a common sense to use it... until you are from Austin where even radio...
Viewing 14 posts - 1 through 13 (of 13 total)
You must be logged in to reply to this topic. Login to reply