I am writing a python regex that matches only string that consists of letters, digits and one or more question marks.

For example, regex1: ^[A-Za-z0-9?] $ returns strings with or without ?

I want a regex2 that matches expressions such as ABC123?A, 1AB?CA?, ?2ABCD, ???, 123? but not ABC123, ABC.?1D1, ABC(a)?1d on mysql, I did that and it works:

select *
from (
select * from norm_prod.skill_patterns
where pattern REGEXP '^[A-Za-z0-9?] $') AS XXX
where XXX.pattern not REGEXP '^[A-Za-z0-9] $'

CodePudding user response：

How about something like this:

^(?=.*\?)[a-zA-Z0-9\?] $

As you can see here at regex101.com

Explanation

The (?=.*\?) is a positive lookahead that tells the regex that the start of the match should be followed by 0 or more characters and then a ? - i.e., there should be a ? somewhere in the match.
The [a-zA-Z0-9\?] matches one-or-more occurrences of the characters given in the character class i.e. a-z, A-Z and digits from 0-9, and the question mark ?.

Altogether, the regex first checks if there is a question mark somewhere in the string to be matched. If yes, then it matches the characters mentioned above. If either the ? is not present, or there is some foreign character, then the string is not matched.

CodePudding user response：

You can validate an alphanumeric string with one or more question marks using

where pattern REGEXP '^[A-Za-z0-9]*([?][A-Za-z0-9]*) $'

In Python:

re.search(r'^[A-Za-z0-9]*(?:\?[A-Za-z0-9]*) $', text)

See the regex demo.

Details:

^ - start of string
[A-Za-z0-9]* - zero or more letters or digits
([?][A-Za-z0-9]*) - one or more repetitions of a ? char and then zero or more letters or digits
$ - end of string.

If you plan to apply this to any Unicode string, consider using POSIX character classes:

where pattern REGEXP '^[[:alnum:]]*([?][[:alnum:]]*) $'

where [[:alnum:]] matches any letters and digits. In Python:

re.search(r'^[^\W_]*(?:\?[^\W_]*) $', text)

In Python, all shorthand character classes are Unicode aware by default, and the [^\W_] pattern is a \w (that matches letters, digits, connector punctuation) with _ subtracted from it.

CodePudding user response：

If there should be at least a single question mark present using MySQL or Python:

^[A-Za-z0-9]*\?[A-Za-z0-9?]*$

Explanation

^ Start of string
[A-Za-z0-9]* Match optional chars A-Z a-z 0-9
\? Match a question mark
[A-Za-z0-9]* Match optional chars A-Z a-z 0-9 or ?
$ End of string

See a regex demo.

In MySQL double escape the backslash like:

REGEXP '^[A-Za-z0-9]*\\?[A-Za-z0-9?]*$'