January 12, 2013 at 4:34 pm
I have a table where a certain column contains a status id as an integer. I want to sort on this field, but I want to sort in a specific order. Ie, instead of 1,2,3,4,5 I want the order returning as 3,2,4,5,1. That is, all records with statusid 3 first, then 2, etc.
The only way I can think of is something like this:
SELECT ... FROM ... WHERE ... ORDER BY CASE WHEN statusId = 3 THEN 0 WHEN stautsId = 2 THEN 1 ... ETC ...
But I'm wondering if there's another way thats more readable that I've overlooked. I really don't want to do multiple DB hits unless its going to be quicker (queries done via php).
January 12, 2013 at 4:38 pm
Another way...
ORDER BY CHARINDEX(statusId,'32451')
MM
select geometry::STGeomFromWKB(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anuary 12, 2013 at 5:04 pm
Thanks, that looks great. However, the number can become 2 decimal places, so I thought something like:
select ('S'+CAST ([statusId] AS VARCHAR(255))) as sid from tbl_xenix
order by CHARINDEX(sid,'S5S4')
but it says 'sid' is an invalid column, despite the fact its in the column header if I miss out the 'order by' !
Any clues?
January 13, 2013 at 1:39 am
select x.NewSid, sid
from tbl_xenix
CROSS APPLY (SELECT NewSid = ('S'+CAST ([statusId] AS VARCHAR(255)))) x
order by CHARINDEX(x.NewSid,'S5S4')
For better assistance in answering your questions, please read this[/url].
Hidden RBAR: Triangular Joins[/url] / The "Numbers" or "Tally" Table: What it is and how it replaces a loop[/url] Jeff Moden[/url]
January 13, 2013 at 2:46 am
You should do it another way. When normalizing the database there should be a table containg all available statusids. Your data table should then have a foreignn key to the statusid-table. When doing so you can add a column to the statusid-table where you define the sort order.
Something like this:
create table dbo.domStatus(
StatusId int not null,
SortOrder int not null
primary key ( StatusId )
);
create table dbo.SomeDataTable(
Col1 varchar(50),
Col2 varchar(50),
StatusId int not null
);
alter table dbo.SomeDataTable add constraint fk_SomeDataTable_domStatus
foreign key ( StatusId ) references dbo.domStatus ( StatusId );
select ...
from dbo.SomeDataTable d inner join
dbo.domStatus s on
s.StatusId = d.StatusId
ORDER BY s.SortOrder
January 13, 2013 at 11:06 am
Thanks, I actually do have such a table (it translates status id to name and description), so its pretty easy to just add another column with the sort order. Thanks for that, it completely slipped my mind to do it like that.
Actually, I have several web pages, and currently the list of ids for each page is in the PHP file. I think I'll move it into the DB, it'll make it easier to change.
I wish there was a way of thanking/voting users on this forum.
January 13, 2013 at 12:04 pm
steve 50602 (1/13/2013)
Thanks, I actually do have such a table (it translates status id to name and description), so its pretty easy to just add another column with the sort order. Thanks for that, it completely slipped my mind to do it like that.Actually, I have several web pages, and currently the list of ids for each page is in the PHP file. I think I'll move it into the DB, it'll make it easier to change.
I wish there was a way of thanking/voting users on this forum.
I wish there was a way of thanking/voting users on this forum. << You just did.
January 13, 2013 at 3:39 pm
WolfgangE (1/13/2013)
You should do it another way. When normalizing the database there should be a table containg all available statusids. Your data table should then have a foreignn key to the statusid-table. When doing so you can add a column to the statusid-table where you define the sort order.Something like this:
create table dbo.domStatus(
StatusId int not null,
SortOrder int not null
primary key ( StatusId )
);
create table dbo.SomeDataTable(
Col1 varchar(50),
Col2 varchar(50),
StatusId int not null
);
alter table dbo.SomeDataTable add constraint fk_SomeDataTable_domStatus
foreign key ( StatusId ) references dbo.domStatus ( StatusId );
select ...
from dbo.SomeDataTable d inner join
dbo.domStatus s on
s.StatusId = d.StatusId
ORDER BY s.SortOrder
While that may be a good idea for any incredibly simple system, what do you do when you want to order things differently?
Add more columns to the domStatus table? No thanks.
What about when you want the end user to choose a sort order? Add all possible combinations to your domStatus table? No thanks.
There are times when IMHO you can take "Normalisation" too far and this may be one of them, don't you think? Unless you can definitely say that anything with a status will always need to be ORDERED exactly the same way, I would steer clear of putting that sort of metadata into the database design.
If you use something simple like the CHARINDEX example, you can provide flexibility at very little cost to design.
But this is just my opinion and I do appreciate the alternative view put forward by Wolfgang 🙂
MM
select geometry::STGeomFromWKB(0x0106000000020000000103000000010000000B0000001000000000000840000000000000003DD8CCCCCCCCCC0840000000000000003DD8CCCCCCCCCC08408014AE47E17AFC3F040000000000104000CDCCCCCCCCEC3F9C999999999913408014AE47E17AFC3F9C99999999991340000000000000003D0000000000001440000000000000003D000000000000144000000000000000400400000000001040000000000000F03F100000000000084000000000000000401000000000000840000000000000003D0103000000010000000B000000000000000000143D000000000000003D009E99999999B93F000000000000003D009E99999999B93F8014AE47E17AFC3F400000000000F03F00CDCCCCCCCCEC3FA06666666666FE3F8014AE47E17AFC3FA06666666666FE3F000000000000003D1800000000000040000000000000003D18000000000000400000000000000040400000000000F03F000000000000F03F000000000000143D0000000000000040000000000000143D000000000000003D, 0);January 13, 2013 at 4:08 pm
I completely agree with you, but in this case, the sort order will not be changeable by the user (ever), and there will only ever be one sort order. If the order is ever requested to be changed, then its a simple method of updating a table rather than digging around in the PHP code.
If I wanted the sort order changeable via the website, then I would definitely use something like CHARINDEX, as it seems a pretty flexible solution. I'd probably just put the concatenation and CAST inside the CHARINDEX function, rather than using a CROSS APPLY.
January 14, 2013 at 12:43 am
I cannot agree to that and in my opinion there is just one correct answer to this: it depends!
What do you call an "incredibly simple system"? How do you call other systems? And how do know all the background about the one requirment of some special sorting?
The way I have shown works perfekt for some requirements, e.g. sorting the items for cash desk accountings in meaningful way that is standard for all 400 shops.
On the other hand I have never seen a website where a user can define how single values have to be sorted. I only have seen the possibility of giving a column name for sorting, ascending or descending.
The disadvantages using the cross-apply-method I see are performance and maintenance:
If the number of rows of the outer select increases the query will slow down as the apply-operator does the subselect for any outer row.
And what do you do when you have to change the sort logic or if there is a new statusid? In how many procedures will you have to change your code? Will you even find them all? A simple update of 1 table with a column for sort order seems to be much better.
But as I said: it depends.
Giving the user the possibility to choose a column for sorting or even let him decide, which values within a column are sorted as the user wants might be difficult. I wonder if this would even be part of the database procedure at all. On huge systems there might be thougts like "how many execution plan for one query do I want in my procedure cache?" or "will my execution plans be cached and reused?" Maybe it's more efficient doing the query without sorting on database level and sort the data in the frontend-programming.
Yes, there are many aspects. Depending on the background of the requirement you will have to find an appropriate solution.
January 14, 2013 at 2:47 am
steve 50602 (1/13/2013)
If I wanted the sort order changeable via the website, then I would definitely use something like CHARINDEX, as it seems a pretty flexible solution. I'd probably just put the concatenation and CAST inside the CHARINDEX function, rather than using a CROSS APPLY.
Might be but definitely not from query performance perspective.as it doesnt get any help when you put any function on column
like WHERE left (column_name,1) = 'L' ORDER BY CHARINDEX(statusId,'32451')
-------Bhuvnesh----------
I work only to learn Sql Server...though my company pays me for getting their stuff done;-)
January 14, 2013 at 5:12 am
It's worth mentioning that having a function like that on the ORDER BY can prevent index use. Depending on the size of the data set and the complexity of the rest of the query, while that function provides flexibility, it could seriously kill performance. A structural solution such as having a sorting column is a fairly standard approach and not (in my, admittedly, not worth anything more than anyone else's, opinion) an example of over-normalization.
"The credit belongs to the man who is actually in the arena, whose face is marred by dust and sweat and blood"
- Theodore Roosevelt
Author of:
SQL Server Execution Plans
SQL Server Query Performance Tuning
January 14, 2013 at 5:39 am
WolfgangE (1/14/2013)
....The disadvantages using the cross-apply-method I see are performance and maintenance:
If the number of rows of the outer select increases the query will slow down as the apply-operator does the subselect for any outer row....
Using a single CROSS APPLY valued to the row (i.e. with no table reference) to replace a single calculation referenced in two or more parts of the same query is quite likely to improve performance. Performance change, a very slight change in either direction, will scale in a linear manner with rowcount. It's not a subselect. It's most likely to show in the plan as a compute scalar.
For better assistance in answering your questions, please read this[/url].
Hidden RBAR: Triangular Joins[/url] / The "Numbers" or "Tally" Table: What it is and how it replaces a loop[/url] Jeff Moden[/url]
January 15, 2013 at 10:08 am
Here's another alternative you could try:
IF OBJECT_ID('tempdb..#TempTable') IS NOT NULL
DROP TABLE #TempTable
CREATE TABLE #TempTable (
[ID] [int] IDENTITY(1,1) NOT NULL,
[FirstName] [nvarchar](50) NULL,
[LastName] [nvarchar](50) NULL,
[StatusID] [int] NULL,
PRIMARY KEY (ID))
INSERT INTO #TempTable
SELECT 'George','Washington','1'
UNION ALL SELECT 'John','Adams','2'
UNION ALL SELECT 'Thomas','Jefferson','3'
UNION ALL SELECT 'James','Madison','4'
UNION ALL SELECT 'James','Monroe','5'
UNION ALL SELECT 'Andrew','Jackson','6'
UNION ALL SELECT 'John Quincy','Adams','7'
--Original order
SELECT * FROM #TempTable ORDER BY ID
--Order by this string order
DECLARE @SortOrder VARCHAR(50)
SET @SortOrder = '3,2,4,5,1,6,7'
SELECT
ID
,FirstName
,LastName
,StatusID
FROM
#TempTable AS tt
INNER JOIN
dbo.DelimitedSplit8K(@SortOrder,',') AS dsk
ON tt.StatusID = dsk.Item
For those who don't have the DelimitedSplit8K function already:
CREATE FUNCTION [dbo].[DelimitedSplit8K]
(
@pString VARCHAR(8000)
,@pDelimiter CHAR(1)
)
RETURNS TABLE
WITH SCHEMABINDING
AS
RETURN
WITH E1(N)
AS (
SELECT
1
UNION ALL
SELECT
1
UNION ALL
SELECT
1
UNION ALL
SELECT
1
UNION ALL
SELECT
1
UNION ALL
SELECT
1
UNION ALL
SELECT
1
UNION ALL
SELECT
1
UNION ALL
SELECT
1
UNION ALL
SELECT
1
),--10E+1 or 10 rows
E2(N)
AS (
SELECT
1
FROM
E1 a
,E1 b
),--10E+2 or 100 rows
E4(N)
AS (
SELECT
1
FROM
E2 a
,E2 b
),--10E+4 or 10,000 rows max
cteTally(N)
AS (
SELECT TOP (ISNULL(DATALENGTH(@pString),0))
ROW_NUMBER() OVER (ORDER BY (SELECT NULL))
FROM E4
),
cteStart(N1)
AS (
SELECT
1
UNION ALL
SELECT
t.N + 1
FROM
cteTally t
WHERE
SUBSTRING(@pString,t.N,1) = @pDelimiter
),
cteLen(N1,L1)
AS (
SELECT
s.N1
,ISNULL(NULLIF(CHARINDEX(@pDelimiter,@pString,s.N1),0) - s.N1,8000)
FROM
cteStart s
)
SELECT
ItemNumber = ROW_NUMBER() OVER (ORDER BY l.N1)
,Item = SUBSTRING(@pString,l.N1,l.L1)
FROM
cteLen l ;
GO
Viewing 14 posts - 1 through 13 (of 13 total)
You must be logged in to reply to this topic. Login to reply