How do I remove extended ASCII characters from a string in T-SQL?

OK, give this a try. It seems the same issue they have. Anyway you need to modify it based on your requirements.

CREATE FUNCTION RemoveNonASCII 
(
    @nstring nvarchar(255)
)
RETURNS varchar(255)
AS
BEGIN

    DECLARE @Result varchar(255)
    SET @Result = ''

    DECLARE @nchar nvarchar(1)
    DECLARE @position int

    SET @position = 1
    WHILE @position <= LEN(@nstring)
    BEGIN
        SET @nchar = SUBSTRING(@nstring, @position, 1)
        --Unicode & ASCII are the same from 1 to 255.
        --Only Unicode goes beyond 255
        --0 to 31 are non-printable characters
        IF UNICODE(@nchar) between 32 and 255
            SET @Result = @Result + @nchar
        SET @position = @position + 1
    END

    RETURN @Result

END
GO

Check it out at SqlServerCentral


The accepted answer is using a loop which should be avoided...

My solution is completely inlineable, it's easy to create an UDF (or maybe even better: an inline TVF) from this.

The idea: Create a set of running numbers (here it's limited with the count of objects in sys.objects, but there are tons of example how to create a numbers tally on the fly). In the second CTE the strings are splitted to single characters. The final select comes back with the cleaned string.

DECLARE @tbl TABLE(ID INT IDENTITY, EvilString NVARCHAR(100));
INSERT INTO @tbl(EvilString) VALUES('ËËËËeeeeËËËË'),('ËaËËbËeeeeËËËcË');

WITH RunningNumbers AS
(
    SELECT ROW_NUMBER() OVER(ORDER BY (SELECT NULL)) AS Nmbr
    FROM sys.objects
)
,SingleChars AS
(
    SELECT tbl.ID,rn.Nmbr,SUBSTRING(tbl.EvilString,rn.Nmbr,1) AS Chr
    FROM @tbl AS tbl
    CROSS APPLY (SELECT TOP(LEN(tbl.EvilString)) Nmbr FROM RunningNumbers) AS rn 
)
SELECT ID,EvilString
      ,(
        SELECT '' + Chr 
        FROM SingleChars AS sc
        WHERE sc.ID=tbl.ID AND ASCII(Chr)<128
        ORDER BY sc.Nmbr
        FOR XML PATH('')
      ) AS GoodString
FROM @tbl As tbl

The result

1   ËËËËeeeeËËËË    eeee
2   ËaËËbËeeeeËËËcË abeeeec

Here is another answer from me where this approach is used to replace all special characters with secure characters to get plain latin