I want to validate that the user has entered a comma separated list of words only using regex, will this work/is there a better way:
$match = "#^([^0-9 A-z])(,\s|$))+$#";
This is not for parsing, as I will use explode for that, it is merely to validate that the user has correctly understood that values should be separated with commas.
I don't know what the separate values should look like, but perhaps this is something that can help you:
$value = '[0-9 A-Z]+';
$match = "~^$value(,$value)*$~i";
When you know what your values should look like you can change $value.
Taken this comment from the question
I was trying to disallow 0-9 but allow a-Z, because, as I said it should be a list of words only, no spaces, just a list of single words.
Just change $value
$value = '[A-Z]+';
Note, that I use the i-modifier in $match, that will include all lowercase letters too.
However, "real" csv allows a little bit more than this, for example you can put quotes around every value. I would recommend that you parse the line and then test every single value
$values = str_getcsv($line);
$valid = true;
foreach ($values as $value) {
$valid = $valid && isValid($value);
}
[0-9 A-z]) only covers a little subset of what is possible. Additional more important than the syntactical correctness is the semantically one, means: You should usually also ensure the number of values per line (and so on).
" , " passes this test$value it is valid (results in array(' ',' ')). As mentioned in the question I don't really understand, when a value should be treated as "valid", thus I decided to outsource the "value"-definition, so everybody can change $value like they want. Its also generally valid for csv (independent from this regex, see comments under k102s question).your $match gives an error.
this
$str = 'sdfbdf,wefwef,323r,dfvdfv';
$match = "/[\S\,]+\S/";
preg_match($match,$str,$m);
can work, but why don't you use explode?
$str = ',323r,,,dfvdfv';';
explode or str_getcsv as you said.I think this is what you're looking for:
'#\G(?:[A-Za-z]+(?:\s*,\s*|$))+$#'
\G anchors the match either to the beginning of the string or the position where the previous match ended. That ensures that the regex doesn't skip over any invalid characters as it tries to match each word. For example, given this string:
'foo,&&bar'
It will report failure because it can't start the second match immediately after the comma.
Also, notice the character class: [A-Za-z]. You used [A-z] in your regex and [a-Z] in a comment, both of which are invalid (for this purpose, anyway). They may have been mere typos, but watch out for them nonetheless. Those typos could could end up causing subtle and/or serious bugs.
EDIT: \G isn't universally supported, so check before using it in any another regex flavors.
I suggest, if you want to test regular expressions, to use kiki.
About your regex: it won't work because of unmatched parentheses. Some more things to take into account:
Besides, what's the point of validating whether or not this is a comma-seperated list?
0-9A-zbefore the first comma (for example)