Becs character set regex

Question

I am working on a project where i have to validate the BECS characters. Bulk Electronic Clearing System (BECS) only allowed the following characters.

BECS Character set

Type                        Description
Numeric                     0 to 9
Alphabetic                  Uppercase A to Z and Lowercase a to z
+                           Plus sign
-                           Minus sign or hyphen
@                           At sign
SP                          Blank space
$                           Dollar sign
!                           Exclamation mark
%                           Percentage sign
&                           Ampersand
(                           Left Parenthesis
)                           Right Parenthesis
*                           Asterik
.                           Period or decimal point
/                           Solidus (slash)
#                           Number Sign (Pound or Hash)
=                           Equal Sign
:                           Colon
;                           Semicolon
?                           Question mark
,                           Comma
’                           Apostrophe
[                           Left square bracket
]                           Right square bracket
_                           Low line (underscore)
^                           Circumflex

I have tried the following but not working:

preg_match("/^[A-Za-z0-9^_[]',?;:=#/.*()&%!$ @+-]+$/", $string);

I also recommend using this $string = iconv("UTF-8", "ASCII//TRANSLIT", $string); — Yevgeniy Afanasyev, Dec 18 '20 at 06:42

b0ne · Answer 1 · 2016-08-15T08:13:03.507

6

Instead of worrying about escaping manually use preg_quote

The code would then be preg_match("/^[A-Za-z0-9".preg_quote("^_[]',?;:=#/.*()&%!$ @+-", "/")."]+$/", $string);

edited Aug 15 '16 at 08:13

answered Aug 15 '16 at 08:07

b0ne

653
3
10

`preg_quote` is for escaping literal strings, not characters inside a character class. E.g. in this case, it returns `A\-Za\-z0\-9\^_\[\]',\?;\:\=#\/\.\*&%\!\$ @\+\-`, which is wrong – Mariano Aug 15 '16 at 08:11
I've given this an upvote because it's a good suggestion generally. However regex escaping within a character range (ie within square brackets) has different quoting rules, so preg_quote probably won't be quite right in this context. – Simba Aug 15 '16 at 08:12
1

Thanks for correcting it, upvoted. It will escape more characters than what's needed, but this is safe as long as you don't pass character ranges to `preg_quote`. – Mariano Aug 15 '16 at 08:21
If you want to make it even more cleaner then you could define your regex in two variables: one for range and one for other matches and then escape only the matches variable. This is not necessary if you're only using that regex one time – b0ne Aug 15 '16 at 08:23

Mariano · Accepted Answer · 2016-08-15T08:33:09.047

4

Inside character classes, you don't need to escape most of the metacharacters.

/^[A-Za-z0-9^_[\]',?;:=#\/.*()&%!$ @+-]+$/

] is escaped to prevent it from closing the character class
/ needs to be escaped because we're using it as regex delimiter
- doesn't need to be escaped because it's the last character in the class
^ doesn't need to be escaped because it's not the first character in the class

Or, if you want a shorter expression, the following regex covers the same range:

/^[ !#-;=?-[\]^_a-z]+$/

edited Aug 15 '16 at 08:33

answered Aug 15 '16 at 08:01

Mariano

6,423
4
31
47

Thanks you very much. – Anam Aug 15 '16 at 08:05
This method also works but there is a much easier and cleaner way. Take a look at me answer bellow – b0ne Aug 15 '16 at 08:08
I prefer to know and explain why each metacharacter needs or doesn't need to be escaped. – Mariano Aug 15 '16 at 08:36

Jean-François Fabre · Answer 3 · 2016-08-15T10:26:23.260

1

you have to escape most of the chars in your string as they have special meaning in regular expressions

(of course you were right not to escape the leading ^ and the trailing $ as they indicate that no other character can be in the line):

preg_match("/^[A-Za-z0-9\\^_\\[\\]',\\?;:=#/\\.\\*\\(\\)\\&\\%!\\$ @\\+\\-]+$/", $string);

For the record a list of the BECS allowed chars vs regular expression chars:

Type                        Description
Numeric                     0 to 9
Alphabetic                  Uppercase A to Z and Lowercase a to z
+                           Plus sign: means 1 or more
-                           Minus sign or hyphen: used for char range
$                           Dollar sign: end of line
(                           Left Parenthesis: starts a group
)                           Right Parenthesis: ends a group
*                           Asterik: 0 or more
.                           Period or decimal point: any char (not in a range, would work)
?                           Question mark: 0 or 1
[                           Left square bracket: starts a range
]                           Right square bracket: ends a range
^                           Circumflex: not/line start

edited Aug 15 '16 at 10:26

answered Aug 15 '16 at 07:46

Jean-François Fabre

137,073
23
153
219

2

Its not working. `preg_match(): Unknown modifier '\' on line number 5` – Anam Aug 15 '16 at 07:57
added more backslashes. Seeing other answers, they seem more appriopriate, though. – Jean-François Fabre Aug 15 '16 at 09:00
1

Still, you can't have an odd number of backslashes. Also, I'm not a big fan of the "If you don't know the rules, escape everything"... – Thomas Ayoub Aug 15 '16 at 09:01
I'm editing the answer. Still for the triple backslash before $ I followed this answer: http://stackoverflow.com/questions/9716443/how-to-escape-dollar-sign-in-a-string-using-perl-regex – Jean-François Fabre Aug 15 '16 at 09:06
Your link is for a perl program, not a PHP one. Running your code, I get a `PHP Warning: preg_match(): Unknown modifier '\' in prog.php on line 4` because the `/` is not escaped. Also, it would have been a lot easier to use `"/^[\w^[\\]',?;:=#\\/.*()&%!$ @+-]+$/"` – Thomas Ayoub Aug 15 '16 at 10:22

Becs character set regex

3 Answers3