[SOLVED] Matching a price.

[play_]

Hi.

 

Looking for a regular expression that would match a price. (without $).

 

So it would match:

• 10
  
• 10.99
  
• 1,099
  
• 1,099.49
  

 

So i guess the rule is, has to start with a number, has to end with a number, and can have one comma and/or one period in the middle

[Garethp]

~^([0-9]+(?:\.|,)?[0-9]+?$~

[nrg_alpha]

~^([0-9]+(?:\.|,)?[0-9]+?$~

 

Of course, that pattern would fail if you encounter a value like 1,345,675.90 (since yours only allows a single comma or period).

 

My example:

setlocale(LC_CTYPE, 'C'); // ensure \d only matches 0-9
$price = array('10', '10.99', '1,099', '1,099.98', '1,34,5.68', '123.56.7', '3,454,579.9', '1,665.908');
foreach($price as $val){
    echo (preg_match('#^\d+(?:,\d{3})*(\.)?(?(1)\d{1,2})$#', $val))? $val . " is valid!<br />\n" : $val . " is NOT valid!<br />\n";
}

 

@play_, notice the final quantifying interval {1,2}... this would allow one or two digits after the decimal (if decimal is found). If you only want two digits after the decimal, simply change that interval to {2}

[Garethp]

Why not just ~^([0-9]{1,3})((,|\.)[0-9]{3,3})*?(\.[0-9]{1.2}$~

 

That way you have 1 - 3 numbers in the front, and then possible a comma or dot followed by three numbers, which is a repitive optional, with an option two decimal number at the end?

 

When I say why not, I'm not trying to one up you, I'm just trying to see if there's something wrong with my RegEx, I'm still kinda new at it

[nrg_alpha]

Why not just ~^([0-9]{1,3})((,|\.)[0-9]{3,3})*?(\.[0-9]{1.2}$~

 

That way you have 1 - 3 numbers in the front, and then possible a comma or dot followed by three numbers, which is a repetitive optional, with an option two decimal number at the end?

 

When I say why not, I'm not trying to one up you, I'm just trying to see if there's something wrong with my RegEx, I'm still kinda new at it

 

There are a few issues with your pattern. It looks for a digit (1-3 times), followed by a comma or dot, followed by a digit, 3 times (I'll get to your additional errors in a second).. so if I stop right there, that could in itself match 3.543,678.432,788  see the problem? The cluster of 3 digits can alternate between commas and dots..

 

Additional issues with this pattern include the following..

 

((,|\.)[0-9]{3,3})*? Since you are using zero or more times, you don't need the ? to make it optional, as it is already optional (because it allows zero times)

 

Also, keep in mind that every set of parenthesis creates a capture.. since we are looking for a format, we don't care about doing specific captures.. so when you have something like (,|\.), this is getting stored into an internal variable.. when dealing with stuff we don't want to capture, we can resort to non-capture groups, which take on the format (?: .... ). In my pattern in my previous post, I make use of only 1 capture (the decimal), and if that decimal is there, then it must follow 1 or 2 digits. Earlier in my pattern, you'll notice (?:,\d{3})* which groups a comma followed by 3 digits, all of which is zero or more times.. but I don't care about storing this into an internal value.. so I use (?:,\d{3}) instead of (,\d{3}).

 

[0-9]{3,3} intervals like that {3,3} can simply be stated as {3}. When there is two values involved, the first is a minimum, the second a maximum.. so in this case, you are saying 3 minimum, 3 maximum.. by only stating a single value, this says, I want an x amount only (so {3} means look for something 3 consecutive times).

 

(\.[0-9]{1.2} there are two issues with this... a) there is no closing parenthesis.. b), your interval is using a dot instead of a comma to separate min and max values.

 

[nrg_alpha]

Granted, I did realize I botched the beginning of my pattern because of \d+... so my pattern should be more like:

 

#^\d{1,3}(?:,\d{3})*(\.)?(?(1)\d{2})$#

[play_]

Thanks to both for replying.

 

neg_alpha, i am using yours, without setlocale function. seems to work fine without it.

[nrg_alpha]

Yeah, the setlocale was just incase.. (odds are, you won't have to deal with any additional stuff like exponents.. just a habit I have when dealing with short hand character classes), so it shouldn't be an issue without it. Sorry I botched my initial pattern... its 2:30am here.. Im tired. 

[play_]

it is late

 

and thanks again. works beautifully.

[Garethp]

Why not just ~^([0-9]{1,3})((,|\.)[0-9]{3,3})*?(\.[0-9]{1.2}$~

 

That way you have 1 - 3 numbers in the front, and then possible a comma or dot followed by three numbers, which is a repetitive optional, with an option two decimal number at the end?

 

When I say why not, I'm not trying to one up you, I'm just trying to see if there's something wrong with my RegEx, I'm still kinda new at it

 

There are a few issues with your pattern. It looks for a digit (1-3 times), followed by a comma or dot, followed by a digit, 3 times (I'll get to your additional errors in a second).. so if I stop right there, that could in itself match 3.543,678.432,788  see the problem? The cluster of 3 digits can alternate between commas and dots..

 

Additional issues with this pattern include the following..

 

((,|\.)[0-9]{3,3})*? Since you are using zero or more times, you don't need the ? to make it optional, as it is already optional (because it allows zero times)

 

Also, keep in mind that every set of parenthesis creates a capture.. since we are looking for a format, we don't care about doing specific captures.. so when you have something like (,|\.), this is getting stored into an internal variable.. when dealing with stuff we don't want to capture, we can resort to non-capture groups, which take on the format (?: .... ). In my pattern in my previous post, I make use of only 1 capture (the decimal), and if that decimal is there, then it must follow 1 or 2 digits. Earlier in my pattern, you'll notice (?:,\d{3})* which groups a comma followed by 3 digits, all of which is zero or more times.. but I don't care about storing this into an internal value.. so I use (?:,\d{3}) instead of (,\d{3}).

 

[0-9]{3,3} intervals like that {3,3} can simply be stated as {3}. When there is two values involved, the first is a minimum, the second a maximum.. so in this case, you are saying 3 minimum, 3 maximum.. by only stating a single value, this says, I want an x amount only (so {3} means look for something 3 consecutive times).

 

(\.[0-9]{1.2} there are two issues with this... a) there is no closing parenthesis.. b), your interval is using a dot instead of a comma to separate min and max values.

 

So how about

 

~^((?:[0-9]{1,3})(?:,[0-9]{3})*(?:\.[0-9]{1,2}))$~

 

Would that work? And as for capturing the data, he just said match, didn't he? Still, my above one should match the price as a number if I wrote it properly

[nrg_alpha]

So how about

 

~^((?:[0-9]{1,3})(?:,[0-9]{3})*(?:\.[0-9]{1,2}))$~

 

Would that work? And as for capturing the data, he just said match, didn't he? Still, my above one should match the price as a number if I wrote it properly

 

Would it work? You could always try it yourself on an array of possible combinations of dollar figures (as I have done with my example).

It would almost work. The first and second groups work... (?:[0-9]{1,3})(?:,[0-9]{3})* but the last part - (?:\.[0-9]{1,2}) doesn't, because after the initial 3 digits in the number, whether the middle part exists or not [referring to (?:,[0-9]{3})*], the final part requires a decimal, followed by one or two digits. That last part could have the ? quantifier.

 

While you have done better with non capturing, you have encased the entire pattern in a capture (the very first and outer last parenthesis is not necessary). Also, you don't need to put a non capturing group on (?:[0-9]{1,3}) at the start, as this sequence is already required.. so one possible version based on yours could be (untested):

 

~^[0-9]{1,3}(?:,[0-9]{3})*(?:\.[0-9]{1,2})?$~

 

That way, after the first initial 3 digits, both second and last chunks are optional.