如何从 T-SQL 中的字符串中删除扩展 ASCII 字符？
声明：本页面是StackOverFlow热门问题的中英对照翻译，遵循CC BY-SA 4.0协议，如果您需要使用它，必须同样遵循CC BY-SA许可，注明原文地址和作者信息，同时你必须将它归于原作者（不是我）：StackOverFlow
原文地址: http://stackoverflow.com/questions/15259622/
Warning: these are provided under cc-by-sa 4.0 license. You are free to use/share it, But you must attribute it to the original authors (not me):
StackOverFlow
How do I remove extended ASCII characters from a string in T-SQL?
提问by Codeman
I need to filter out (remove) extended ASCII characters from a SELECT statement in T-SQL.
我需要从 T-SQL 的 SELECT 语句中过滤掉（删除）扩展的 ASCII 字符。
I'm using a stored procedure to do so.
我正在使用存储过程来这样做。
Expected input:
预期输入：
????eeee????
Expected output:
预期输出：
eeee
All that I've found is for MySQL.
我发现的所有内容都是针对MySQL 的。
I'm using :
我正在使用 ：
Microsoft SQL Server Management Studio 11.0.2100.60
Microsoft .NET Framework 4.0.30319.17929
回答by amit kohan
OK, give this a try. It seems the same issue they have. Anyway you need to modify it based on your requirements.
好的，试试这个。他们似乎有同样的问题。无论如何，您需要根据您的要求对其进行修改。
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
回答by Shnugo
The accepted answer is using a loop which should be avoided...
接受的答案是使用一个应该避免的循环......
My solution is completely inlineable, it's easy to create an UDF (or maybe even better: an inline TVF) from this.
我的解决方案是完全内联的，很容易从中创建 UDF（或者甚至更好：内联 TVF）。
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.
想法：创建一组运行数字（这里受 sys.objects 中对象数量的限制，但有大量示例如何即时创建数字计数）。在第二个 CTE 中，字符串被拆分为单个字符。最终选择返回清洁后的字符串。
DECLARE @tbl TABLE(ID INT IDENTITY, EvilString NVARCHAR(100));
INSERT INTO @tbl(EvilString) VALUES('????eeee????'),('?a??b?eeee???c?');
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
结果
1 ????eeee???? eeee
2 ?a??b?eeee???c? abeeeec
Here is another answerfrom me where this approach is used to replace all specialcharacters with securecharacters to get plain latin
回答by Clicks to Bricks
Just correcting above code (it was cutting DOTs)
只是更正上面的代码（它正在削减 DOT）
CREATE FUNCTION [dbo].[RemoveNonASCII]
(
@nstring nvarchar(255)
)
RETURNS nvarchar(255)
AS
BEGIN
DECLARE @Result nvarchar(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 192 and 198) or (UNICODE(@nchar) between 225 and 230) -- letter A or a with accents
SET @nchar = 'a'
IF (UNICODE(@nchar) between 200 and 203) or (UNICODE(@nchar) between 232 and 235) -- letter E or e with accents
SET @nchar = 'e'
IF (UNICODE(@nchar) between 204 and 207) or (UNICODE(@nchar) between 236 and 239) -- letter I or i with accents
SET @nchar = 'i'
IF (UNICODE(@nchar) between 210 and 214) or (UNICODE(@nchar) between 242 and 246) or (UNICODE(@nchar)=240) -- letter O or o with accents
SET @nchar = 'o'
IF (UNICODE(@nchar) between 217 and 220) or (UNICODE(@nchar) between 249 and 252) -- letter U or u with accents
SET @nchar = 'u'
IF (UNICODE(@nchar)=199) or (UNICODE(@nchar)=231) -- letter ? or ?
SET @nchar = 'c'
IF (UNICODE(@nchar)=209) or (UNICODE(@nchar)=241) -- letter ? or ?
SET @nchar = 'n'
IF (UNICODE(@nchar) between 45 and 46) or (UNICODE(@nchar) between 48 and 57) or (UNICODE(@nchar) between 64 and 90) or (UNICODE(@nchar) = 95) or (UNICODE(@nchar) between 97 and 122)
SET @Result = @Result + @nchar
SET @position = @position + 1
END
set @Result = lower(@Result) -- e-mails in lower case
RETURN @Result
END
回答by AlxSts
Thank you for sharing your code.
感谢您分享您的代码。
I needed something like it, not just for cleaning up e-mail addresses but for general purpose, filtering user web site input before it reaches SAP ERP via integration modules.
我需要类似的东西，不仅用于清理电子邮件地址，而且用于一般目的，在通过集成模块到达 SAP ERP 之前过滤用户网站输入。
Running in Brazil, it must obey language accents...
在巴西运行，它必须遵守语言口音......
Here goes the resulting code.
这是结果代码。
Maybe it can help someone someday, like it did to me.
也许有一天它可以帮助某人，就像对我一样。
IF EXISTS
(
SELECT *
FROM sys.objects
WHERE object_id = OBJECT_ID(N'[dbo].[fnRemoveNonASCII]')
AND type IN (N'FN')
)
DROP FUNCTION dbo.fnRemoveNonASCII
GO
CREATE FUNCTION [dbo].[fnRemoveNonASCII]
(
@nstring nvarchar(MAX)
)
RETURNS nvarchar(MAX)
AS
BEGIN
DECLARE @nchar nvarchar(1) -- individual char in string
DECLARE @nUnicode nvarchar(3) -- ASCII for individual char in string
DECLARE @position int -- subscript to control loop in the string
DECLARE @Result nvarchar(MAX) -- return valus
SET @Result = ''
SET @position = 1
WHILE @position <= LEN(@nstring)
BEGIN
--Unicode & ASCII are the same from 1 to 255.
--Only Unicode goes beyond 255
--0 to 31 are non-printable characters
SET @nchar = SUBSTRING(@nstring, @position, 1)
SET @nUnicode = UNICODE(@nChar)
IF @nUnicode = 10
OR @nUnicode = 13
OR @nUnicode BETWEEN 32 AND 126
OR @nUnicode = 160
OR @nUnicode BETWEEN 192 AND 207
OR @nUnicode BETWEEN 210 AND 213
OR @nUnicode BETWEEN 217 AND 219
OR @nUnicode BETWEEN 224 AND 227
OR @nUnicode BETWEEN 231 AND 234
OR @nUnicode = 236
OR @nUnicode = 237
OR @nUnicode BETWEEN 242 AND 245
OR @nUnicode = 247
OR @nUnicode = 249
OR @nUnicode = 250
SET @Result = @Result + @nchar
ELSE IF @nUnicode = 9 -- TAB
SET @Result = @Result + ' '
ELSE
SET @Result = @Result + ' '
SET @position = @position + 1
END
RETURN @Result
END
/*
---------------------------------------------------------------------------------------------------------------
-- Tabela dos caracteres Unicode/ASCII exportáveis
	 | 	 | %9 = TAB
| 
 | %a = 0A Line Feed
| 
 | %d = 0D Carriage Return
  |   | %20 = <space>
! | ! | %21 = !
" | " | %22 = "
# | # | %23 = #
$ | $ | %24 = $
% | % | %25 = %
& | & | %26 = &
' | ' | %27 = '
( | ( | %28 = (
) | ) | %29 = )
* | * | %2a = *
+ | + | %2b = +
, | , | %2c = ,
- | - | %2d = -
. | . | %2e = .
/ | / | %2f = /
0 | 0 | %30 = 0
1 | 1 | %31 = 1
2 | 2 | %32 = 2
3 | 3 | %33 = 3
4 | 4 | %34 = 4
5 | 5 | %35 = 5
6 | 6 | %36 = 6
7 | 7 | %37 = 7
8 | 8 | %38 = 8
9 | 9 | %39 = 9
: | : | %3a = :
; | ; | %3b = ;
< | < | %3c = <
= | = | %3d = =
> | > | %3e = >
? | ? | %3f = ?
@ | @ | %40 = @
A | A | %41 = A
B | B | %42 = B
C | C | %43 = C
D | D | %44 = D
E | E | %45 = E
F | F | %46 = F
G | G | %47 = G
H | H | %48 = H
I | I | %49 = I
J | J | %4a = J
K | K | %4b = K
L | L | %4c = L
M | M | %4d = M
N | N | %4e = N
O | O | %4f = O
P | P | %50 = P
Q | Q | %51 = Q
R | R | %52 = R
S | S | %53 = S
T | T | %54 = T
U | U | %55 = U
V | V | %56 = V
W | W | %57 = W
X | X | %58 = X
Y | Y | %59 = Y
Z | Z | %5a = Z
[ | [ | %5b = [
\ | \ | %5c = \
] | ] | %5d = ]
^ | ^ | %5e = ^
_ | _ | %5f = _
` | ` | %60 = `
a | a | %61 = a
b | b | %62 = b
c | c | %63 = c
d | d | %64 = d
e | e | %65 = e
f | f | %66 = f
g | g | %67 = g
h | h | %68 = h
i | i | %69 = i
j | j | %6a = j
k | k | %6b = k
l | l | %6c = l
m | m | %6d = m
n | n | %6e = n
o | o | %6f = o
p | p | %70 = p
q | q | %71 = q
r | r | %72 = r
s | s | %73 = s
t | t | %74 = t
u | u | %75 = u
v | v | %76 = v
w | w | %77 = w
x | x | %78 = x
y | y | %79 = y
z | z | %7a = z
{ | { | %7b = {
| | | | %7c = |
} | } | %7d = }
~ | ~ | %7e = ~
  |   | %a0 = <nbsp>
À | À | %c0 = à
Á | Á | %c1 = á
 |  | %c2 = ?
à | à | %c3 = ?
Ä | Ä | %c4 = ?
Å | Å | %c5 = ?
Æ | Æ | %c6 = ?
Ç | Ç | %c7 = ?
È | È | %c8 = è
É | É | %c9 = é
Ê | Ê | %ca = ê
Ë | Ë | %cb = ?
Ì | Ì | %cc = ì
Í | Í | %cd = í
Î | Î | %ce = ?
Ï | Ï | %cf = ?
Ò | Ò | %d2 = ò
Ó | Ó | %d3 = ó
Ô | Ô | %d4 = ?
Õ | Õ | %d5 = ?
Ù | Ù | %d9 = ù
Ú | Ú | %da = ú
Û | Û | %db = ?
à | à | %e0 = à
á | á | %e1 = á
â | â | %e2 = a
ã | ã | %e3 = ?
ç | ç | %e7 = ?
è | è | %e8 = è
é | é | %e9 = é
ê | ê | %ea = ê
ì | ì | %ec = ì
í | í | %ed = í
ò | ò | %f2 = ò
ó | ó | %f3 = ó
ô | ô | %f4 = ?
õ | õ | %f5 = ?
÷ | ÷ | %f7 = ÷
ù | ù | %f9 = ù
ú | ú | %fa = ú
*/
GO
回答by brenkdar
The solution provided by amit kohan works, but was to slow for me. With large string data, the iteration trough each character is not optimal. I've made the following function, with isn't very compact but fast.
amit kohan 提供的解决方案有效，但对我来说很慢。对于大字符串数据，每个字符的迭代槽不是最佳的。我做了以下功能，不是很紧凑但很快。
CREATE FUNCTIONdbo.F_StripLowAscii
(
@Name nvarchar(max)
)
RETURN nvarchar(max) as
BEGIN
DECLARE @Result nvarchar(max)
If @Name IS NULL
RETURN @Name
DECLARE @BlankRange VARCHAR(15)
DECLARE @FoundAt INTEGER
-- ASCII CHAR #0 needs a special treatment
SET @BlankRange = '%[' + CHAR(0) + ']%'
SET @FoundAt = PATINDEX(@BlankRange ,@Name COLLATE SQL_Latin1_General_CP850_Bin)
WHILE @FoundAt > 0
BEGIN
SET @name = left(@name, @FoundAt-1 ) + SUBSTRING(@name, @FoundAt+1, LEN(@Name))
SET @FoundAt = PATINDEX(@BlankRange ,@Name COLLATE SQL_Latin1_General_CP850_Bin)
END
SET @BlankRange = '%[' + CHAR(1)+'-'+CHAR(8) + ']%'
SET @FoundAt = PATINDEX(@BlankRange ,@Name COLLATE SQL_Latin1_General_CP850_Bin)
WHILE @FoundAt > 0
BEGIN
SET @name = Replace(@Name, SUBSTRING(@Name, @FoundAt,1),'')
SET @FoundAt = PATINDEX(@BlankRange ,@Name COLLATE SQL_Latin1_General_CP850_Bin)
END
SET @BlankRange = '%[' + CHAR(11)+'-'+CHAR(12) + ']%'
SET @FoundAt = PATINDEX(@BlankRange ,@Name COLLATE SQL_Latin1_General_CP850_Bin)
WHILE @FoundAt > 0
BEGIN
SET @name = Replace(@Name, SUBSTRING(@Name, @FoundAt,1),'')
SET @FoundAt = PATINDEX(@BlankRange ,@Name COLLATE SQL_Latin1_General_CP850_Bin)
END
SET @BlankRange = '%[' + CHAR(14)+'-'+CHAR(31) + ']%'
SET @FoundAt = PATINDEX(@BlankRange ,@Name COLLATE SQL_Latin1_General_CP850_Bin)
WHILE @FoundAt > 0
BEGIN
SET @name = Replace(@Name, SUBSTRING(@Name, @FoundAt,1),'')
SET @FoundAt = PATINDEX(@BlankRange ,@Name COLLATE SQL_Latin1_General_CP850_Bin)
END
RETURN @Name
END
GO
