How do I sort string alphabetically

Question

I am new to SQL Server as well as to Stack overflow. Please excuse for my mistakes.

Is it possible to sort a value in a column aphabetically? Here is my table

NAME
......
XZYVW
EBACD

My result should look like the below

NAME
......
VWXYZ
ABCDE 

Any idea to get this?

Solution 1

Create a User Defined Scalar Function.

CREATE FUNCTION dbo.Alphaorder (@str VARCHAR(50))
returns VARCHAR(50)
  BEGIN
      DECLARE @len    INT,
              @cnt    INT =1,
              @str1   VARCHAR(50)='',
              @output VARCHAR(50)=''

      SELECT @len = Len(@str)
      WHILE @cnt <= @len
        BEGIN
            SELECT @str1 += Substring(@str, @cnt, 1) + ','

            SET @cnt+=1
        END

      SELECT @str1 = LEFT(@str1, Len(@str1) - 1)

      SELECT @output += Sp_data
      FROM  (SELECT Split.a.value('.', 'VARCHAR(100)') Sp_data
             FROM   (SELECT Cast ('<M>' + Replace(@str1, ',', '</M><M>') + '</M>' AS XML) AS Data) AS A
                    CROSS APPLY Data.nodes ('/M') AS Split(a)) A
      ORDER  BY Sp_data

      RETURN @output
  END

Result:

SELECT dbo.Alphaorder ('XZYVW') --VWXYZ

Solution 2

Even some optimized possibilities are there. Here two function use Bubble sort to sort the char.

CREATE FUNCTION udf_SortString
(
    @string VARCHAR(1000)
)
RETURNS VARCHAR(1000)
AS
BEGIN
    DECLARE @len TINYINT
    DECLARE @i TINYINT
    DECLARE @currentchar CHAR(1)
    DECLARE @swapped BIT
    DECLARE @begin BIT
    DECLARE @nextchar CHAR(1)

    SET @begin = 1
    SET @len = LEN(@string)
    SET @i = 1

    WHILE @begin = 1 OR @swapped = 1
    BEGIN
        SET @swapped = 0
        SET @i = 1
        SET @begin = 0
        WHILE @i <= @len
        BEGIN
            SET @currentchar = SUBSTRING(@string, @i, 1)
            SET @nextchar = SUBSTRING(@string, @i + 1, 1)

            IF @currentchar > @nextchar AND (@nextchar > '')
            BEGIN
                SET @string = dbo.udf_swap(@string, @i, @i + 1)
                SET @swapped = 1
            END

            SET @i = @i + 1
        END
    END

    RETURN(@string)
END

Function 2:

CREATE FUNCTION dbo.udf_Swap
(
    @fullstring VARCHAR(1000),
    @charlocation1 TINYINT,
    @charlocation2 TINYINT
)
RETURNS VARCHAR(1000)
AS
BEGIN
        DECLARE @returnval varchar(1000)
        DECLARE @begin VARCHAR(1000), @middle VARCHAR(1000), @end VARCHAR(1000)
        DECLARE @firstchar CHAR(1), @secondchar CHAR(1), @len INT
        SET @fullstring = LTRIM(RTRIM(@fullstring))
        SET @len = LEN(@fullstring)

    IF @charlocation1 > @len OR @charlocation2 > @len
        SET @returnval = @fullstring
        ELSE
        BEGIN
               SET @firstchar = SUBSTRING(@fullstring, @charlocation1, 1)
               SET @secondchar = SUBSTRING(@fullstring, @charlocation2, 1)
               SET @begin = LEFT(@fullstring, (@charlocation1-1))
               SET @middle = SUBSTRING(@fullstring, @charlocation1+1, (@charlocation2-@charlocation1)-1)
               SET @end = SUBSTRING(@fullstring, @charlocation2+1, @len)
               SET @returnval = @begin + @secondchar + @middle + @firstchar + @end
        END
    RETURN(@returnval)
END

Result:

select dbo.udf_SortString('zxcvbfgrtyuijklm')
--Returns bcfgijklmrtuvxyz

--Reference

--Quick demo here

Solution 3

This will sort the letters alphabetically without using a function. Rewritten for optimization after @Shnugo comment

DECLARE @t table(col varchar(4000))
INSERT @t values('kdjfh'),('zug')

SELECT
  Col, 
  (
    SELECT
      chr
    FROM
      (SELECT TOP(LEN(Col)) 
         SUBSTRING(Col,ROW_NUMBER() OVER(ORDER BY 1/0),1)
       FROM sys.messages) A(Chr)
       ORDER by chr
       FOR XML PATH(''), type).value('.', 'varchar(max)'
      ) SortedCol
FROM
  @t

Result:

col    SortedCol
kdjfh  dfhjk
zug    guz

Comments

[Daniel E.]

This article may help you, techrepublic.com/blog/the-enterprise-cloud/…

[Shnugo]

The voted answers will be very bad performers (WHILE loop, scalar functions...). t-clausen.dk's answer is much better, but will use far to many rows in executions. You might want to have a read here...

[Sean Lange]

Whenver I see things like this I have to wonder if there are multiple values in a single intersection (which violates 1NF) or if this is just a classroom assignment. This kind of thing in the real world is a real challenge.

[Tim Schmelter]

@MohithKm: no, you either have to write a t-sql function yourself, use a .NET method(you can also use .NET in SQL-Server) or refactor your database design. The easiest would be to sort in on client-side with .NET where you want to show it.

[Shnugo]

Hi, this will use far to many rows in any execution... You might want to read this answer. Over there I provide an approach to limit the actual count of rows to the actual count of characters per string. But +1 from my side as this is much better, that the other answers ;-)

[Shnugo]

This will be very slow... If interested you might want to read my comment below the question...

[Shnugo]

This will be very slow... If interested you might want to read my comment below the question...

[t-clausen.dk]

@Shnugo optimized and rewritten