JavaScript Regex to capture repeating part of decimal - javascript

Looking for the best way to take an arbitrary number with potentially repeating decimal part, and discover the repeating part (if present).
Ultimately, I need to decorate the number with overline notation (either with css text-decoration or MathML mline), so I need to know the index of where the repetition begins also.
So I need regex that will get me (or can be used in an algorithm to get) the following results:
1.333 // result: {"pattern": 3, index: 0}
1.5444 // result: {"pattern": 4, index: 1}
1.123123 // result: {"pattern": 123, index: 0}
1.5432121212 // result: {"pattern": 12, index: 4}
1.321 // result: null
1.44212 // result: null
Additional Example (from comments):
1.3333 // result: { "pattern": 3, index: 0}

function getRepetend(num) {
var m = (num+'').match(/\.(\d*?)(\d+?)\2+$/);
return m && {pattern: +m[2], index: m[1].length};
}
It works like this:
First, convert the number to string in order to be able to use regular expressions.
Then, match this regex: /\.(\d*?)(\d+)\2+$/:
\. matches the decimal dot.
(\d*?) matches the digits between the decimal dot and the repetend, and captures the result into backreference number 1.
(\d+?) matches the repetend, and captures it into backreference number 2.
\2+ matches repetitions of the repetend.
$ matches end of string.
Finally, if the match is null (i.e. there is no match), return null.
Otherwise, return an object that contains the repetend (backreference 2) converted to number, and the number of digits between the dot and the repetend (backreference 1).

You could try something like this:
(\d+?)\1+$
http://regex101.com/r/eX8eC3/3
It matched some number of digits and then uses a backreference to try and match the same set immediately afterwards 1 or more times. It's anchored at the end of the string because otherwise it'll be tripped up by, for example:
1.5432121212
It would see the 21 repeating instead of the 12.
Adding ? to the first group to make it non-greedy should fix the problem with 1.3333 as raised by Louis.

You can use this regex with RexExp#exec and use result.index in the resulting object:
var re = /(\d+)\1$/;
var s = '.5439876543212211211';
var result = re.exec( s );
console.log ( result.index );
//=> 14
console.log ( result[1] );
//=> 211
JsFiddle Demo

(.+)(?:\1)+$
Try this.See demo.
http://regex101.com/r/uH3tP3/10

The accepeted answer is OKish as per the given examples in the question are concerned. However if one day you find yourself here what you probably need is exactly what the topic says.
JavaScript Regex to capture repeating part of decimal
So you have the floating part of a string and you want to know if it is repeating or not. The accepted answer fails in practice. You can never guarantee that the string ends when the repeating part ends. Most of the time the string ends with a portion of the repeating part or sometimes *thanks* to the double precision errors jumps to irrelevant figures towards the end. So my suggestion is
/(\d+)\1+(?=\d*$)/g
Now this is not a silver bullet. It's helpful but won't protect you from vampires like 3.1941070707811985 which happens to have no repetend at all. In order to feel it you have to develop deeper mechanisms. However in most cases it's just fine in percieving the repend like in
3.1941070707811985 // 07 which is wrong so prove it later
7.16666666810468 // 666 but reduce it to 6 later
3.00000000000001 // 000000 but reduce it to "" later
0.008928571428571428 // 285714 just fine, do nothing
It is not an easy task to find if the floating part of a decimal has repetends or not in this environment. Most possibly you need to do further processing on the given string and the result of the regex for futher reduction / decision.

Related

Applying currency format using replace and a regular expression

I am trying to understand some code where a number is converted to a currency format. Thus, if you have 16.9 it converts to $16.90. The problem with the code is if you have an amount over $1,000, it just returns $1, an amount over $2,000 returns $2, etc. Amounts in the hundreds show up fine.
Here is the function:
var _formatCurrency = function(amount) {
return "$" + parseFloat(amount).toFixed(2).replace(/(\d)(?=(\d{3})+\.)/g, '$1,')
};
(The reason the semicolon is after the bracket is because this function is in itself a statement in another function. That function is not relevant to this discussion.)
I found out that the person who originally put the code in there found it somewhere but didn't fully understand it and didn't test this particular scenario. I myself have not dealt much with regular expressions. I am not only trying to fix it, but to understand how it is working as it is now.
Here's what I've found out. The code between the backslash after the open parenthesis and the backslash before the g is the pattern. The g means global search. The \d means digit, and the (?=\d{3})+\. appears to mean find 3 digits plus a decimal point. I'm not sure I have that right, though, because if that was correct shouldn't it ignore numbers like 5.4? That works fine. Also, I'm not sure what the '$1,' is for. It looks to me like it is supposed to be placed where the digits are, but wouldn't that change all the numbers to $1? Also, why is there a comma after the 1?
Regarding your comment
I was hoping to just edit the regex so it would work properly.
The regex you are currently using is obviously not working for you so I think you should consider alternatives even if they are not too similar, and
Trying to keep the code change as small as possible
Understandable but sometimes it is better to use a code that is a little bit bigger and MORE READABLE than to go with compact and hieroglyphical.
Back to business:
I'm assuming you are getting a string as an argument and this string is composed only of digits and may or may not have a dot before the last 1 or 2 digts. Something like
//input //intended output
1 $1.00
20 $20.00
34.2 $34.20
23.1 $23.10
62516.16 $62,516.16
15.26 $15.26
4654656 $4,654,656.00
0.3 $0.30
I will let you do a pre-check of (assumed) non-valids like 1. | 2.2. | .6 | 4.8.1 | 4.856 | etc.
Proposed solution:
var _formatCurrency = function(amount) {
amount = "$" + amount.replace(/(\d)(?=(\d{3})+(\.(\d){0,2})*$)/g, '$1,');
if(amount.indexOf('.') === -1)
return amount + '.00';
var decimals = amount.split('.')[1];
return decimals.length < 2 ? amount + '0' : amount;
};
Regex break down:
(\d): Matches one digit. Parentheses group things for referencing when needed.
(?=(\d{3})+(\.(\d){0,2})*$). Now this guy. From end to beginning:
$: Matches the end of the string. This is what allows you to match from the end instead of the beginning which is very handy for adding the commas.
(\.(\d){0,2})*: This part processes the dot and decimals. The \. matches the dot. (\d){0,2} matches 0, 1 or 2 digits (the decimals). The * implies that this whole group can be empty.
?=(\d{3})+: \d{3} matches 3 digits exactly. + means at least one occurrence. Finally ?= matches a group after the main expression without including it in the result. In this case it takes three digits at a time (from the end remember?) and leaves them out of the result for when replacing.
g: Match and replace globally, the whole string.
Replacing with $1,: This is how captured groups are referenced for replacing, in this case the wanted group is number 1. Since the pattern will match every digit in the position 3n+1 (starting from the end or the dot) and catch it in the group number 1 ((\d)), then replacing that catch with $1, will effectively add a comma after each capture.
Try it and please feedback.
Also if you haven't already you should (and SO has not provided me with a format to stress this enough) really really look into this site as suggested by Taplar
The pattern is invalid, and your understanding of the function is incorrect. This function formats a number in a standard US currency, and here is how it works:
The parseFloat() function converts a string value to a decimal number.
The toFixed(2) function rounds the decimal number to 2 digits after the decimal point.
The replace() function is used here to add the thousands spearators (i.e. a comma after every 3 digits). The pattern is incorrect, so here is a suggested fix /(\d)(?=(\d{3})+\.)/g and this is how it works:
The (\d) captures a digit.
The (?=(\d{3})+\.) is called a look-ahead and it ensures that the captured digit above has one set of 3 digits (\d{3}) or more + followed by the decimal point \. after it followed by a decimal point.
The g flag/modifier is to apply the pattern globally, that is on the entire amount.
The replacement $1, replaces the pattern with the first captured group $1, which is in our case the digit (\d) (so technically replacing the digit with itself to make sure we don't lose the digit in the replacement) followed by a comma ,. So like I said, this is just to add the thousands separator.
Here are some tests with the suggested fix. Note that it works fine with numbers and strings:
var _formatCurrency = function(amount) {
return "$" + parseFloat(amount).toFixed(2).replace(/(\d)(?=(\d{3})+\.)/g, '$1,');
};
console.log(_formatCurrency('1'));
console.log(_formatCurrency('100'));
console.log(_formatCurrency('1000'));
console.log(_formatCurrency('1000000.559'));
console.log(_formatCurrency('10000000000.559'));
console.log(_formatCurrency(1));
console.log(_formatCurrency(100));
console.log(_formatCurrency(1000));
console.log(_formatCurrency(1000000.559));
console.log(_formatCurrency(10000000000.559));
Okay, I want to apologize to everyone who answered. I did some further tracing and found out the JSON call which was bringing in the amount did in fact have a comma in it, so it is just parsing that first digit. I was looking in the wrong place in the code when I thought there was no comma in there already. I do appreciate everyone's input and hope you won't think too bad of me for not catching that before this whole exercise. If nothing else, at least I now know how that regex operates so I can make use of it in the future. Now I just have to go about removing that comma.
Have a great day!
Assuming that you are working with USD only, then this should work for you as an alternative to Regular Expressions. I have also included a few tests to verify that it is working properly.
var test1 = '16.9';
var test2 = '2000.5';
var test3 = '300000.23';
var test4 = '3000000.23';
function stringToUSD(inputString) {
const splitValues = inputString.split('.');
const wholeNumber = splitValues[0].split('')
.map(val => parseInt(val))
.reverse()
.map((val, idx, arr) => idx !== 0 && (idx + 1) % 3 === 0 && arr[idx + 1] !== undefined ? `,${val}` : val)
.reverse()
.join('');
return parseFloat(`${wholeNumber}.${splitValues[1]}`).toFixed(2);
}
console.log(stringToUSD(test1));
console.log(stringToUSD(test2));
console.log(stringToUSD(test3));
console.log(stringToUSD(test4));

Javascript regex match returning a string with comma at the end

Just as the title says...i'm trying to parse a string for example
2x + 3y
and i'm trying to get only the coefficients (i.e. 2 and 3)
I first tokenized it with space character as delimiter giving me "2x" "+" "3y"
then i parsed it again to this statement to get only the coefficients
var number = eqTokens[i].match(/(\-)?\d+/);
I tried printing the output but it gave me "2,"
why is it printing like this and how do i fix it? i tried using:
number = number.replace(/[,]/, "");
but this just gives me an error that number.replace is not a function
What's wrong with this?
> "2x + 3y".match(/-?\d+(?=[A-Za-z]+)/g)
[ '2', '3' ]
The above regex would match the numbers only if it's followed by one or more alphabets.
Match is going to return an array of every match. Since you put the optional negative in a parentheses, it's another capture group. That capture group has one term and it's optional, so it'll return an empty match in addition to your actual match.
Input 2x -> Your output: [2,undefined] which prints out as "2,"
Input -2x -> Your output: [2,-]
Remove the parentheses around the negative.
This is just for the sake of explaining why your case is breaking but personally I'd use Avinash's answer.

capture with regex in javascript

I have a string like "ListUI_col_order[01234567][5]". I'd like to capture the two numeric sequences from the string. The last part between the square brackets may contain 2 digits, while the first numeric sequence always contains 8 digits (And the numbers are dynamically changing of course.) Im doing this in javascript and the code for the first part is simple: I get the only 8digit sequence from the string:
var str = $(this).attr('id');
var unique = str.match(/([0-9]){8}/g);
Getting the second part is a bit complicated to me. I cannot simply use:
var column = str.match(/[0-9]{1,2}/g)
Because this will match '01', '23', '45', '67', '5' in our example, It's clear. Although I'm able to get the information what I need as column[4], because the first part always contains 8 digits, but I'd like a nicer way to retrieve the last number.
So I define the contex and I can tell the regex that Im looking for a 1 or 2 digit number which has square brackets directly before and after it:
var column = str.match(/\[[0-9]{1,2}\]/g)
// this will return [5]. which is nearly what I want
So to get Only the numeric data I use parenthesis to capture only the numbers like:
var column = str.match(/\[([0-9]){1,2}\]/g)
// this will result in:
// column[0] = '[5]'
// column[1] = [5]
So my question is how to match the '[5]' but only capture the '5'? I have only the [0-9] between the parenthesis, but this will still capture the square brackets as well
You can get both numbers in one go :
var m = str.match(/\[(\d{8})\]\[(\d{1,2})\]$/)
For your example, this makes ["[01234567][5]", "01234567", "5"]
To get both matches as numbers, you can then do
if (m) return m.slice(1).map(Number)
which builds [1234567, 5]
Unfortunately, JavaScript does not support the lookbehind necessary to do this. In other languages such as PHP, it'd be as simple as /(?<=\[)\d{1,2}(?=\])/, but in JavaScript I am not aware of any way to do this other than use a capturing subpattern as you are here, and getting that index from the result array.
Side-note, it's usually better to put the quantifier inside the capturing group - otherwise you're repeating the group itself, not its contents!

regular expression for finding decimal/float numbers?

i need a regular expression for decimal/float numbers like 12 12.2 1236.32 123.333 and +12.00 or -12.00 or ...123.123... for using in javascript and jQuery.
Thank you.
Optionally match a + or - at the beginning, followed by one or more decimal digits, optional followed by a decimal point and one or more decimal digits util the end of the string:
/^[+-]?\d+(\.\d+)?$/
RegexPal
The right expression should be as followed:
[+-]?([0-9]*[.])?[0-9]+
this apply for:
+1
+1.
+.1
+0.1
1
1.
.1
0.1
Here is Python example:
import re
#print if found
print(bool(re.search(r'[+-]?([0-9]*[.])?[0-9]+', '1.0')))
#print result
print(re.search(r'[+-]?([0-9]*[.])?[0-9]+', '1.0').group(0))
Output:
True
1.0
If you are using mac, you can test on command line:
python -c "import re; print(bool(re.search(r'[+-]?([0-9]*[.])?[0-9]+', '1.0')))"
python -c "import re; print(re.search(r'[+-]?([0-9]*[.])?[0-9]+', '1.0').group(0))"
You can check for text validation and also only one decimal point validation using isNaN
var val = $('#textbox').val();
var floatValues = /[+-]?([0-9]*[.])?[0-9]+/;
if (val.match(floatValues) && !isNaN(val)) {
// your function
}
This is an old post but it was the top search result for "regular expression for floating point" or something like that and doesn't quite answer _my_ question. Since I worked it out I will share my result so the next person who comes across this thread doesn't have to work it out for themselves.
All of the answers thus far accept a leading 0 on numbers with two (or more) digits on the left of the decimal point (e.g. 0123 instead of just 123) This isn't really valid and in some contexts is used to indicate the number is in octal (base-8) rather than the regular decimal (base-10) format.
Also these expressions accept a decimal with no leading zero (.14 instead of 0.14) or without a trailing fractional part (3. instead of 3.0). That is valid in some programing contexts (including JavaScript) but I want to disallow them (because for my purposes those are more likely to be an error than intentional).
Ignoring "scientific notation" like 1.234E7, here is an expression that meets my criteria:
/^((-)?(0|([1-9][0-9]*))(\.[0-9]+)?)$/
or if you really want to accept a leading +, then:
/^((\+|-)?(0|([1-9][0-9]*))(\.[0-9]+)?)$/
I believe that regular expression will perform a strict test for the typical integer or decimal-style floating point number.
When matched:
$1 contains the full number that matched
$2 contains the (possibly empty) leading sign (+/-)
$3 contains the value to the left of the decimal point
$5 contains the value to the right of the decimal point, including the leading .
By "strict" I mean that the number must be the only thing in the string you are testing.
If you want to extract just the float value out of a string that contains other content use this expression:
/((\b|\+|-)(0|([1-9][0-9]*))(\.[0-9]+)?)\b/
Which will find -3.14 in "negative pi is approximately -3.14." or in "(-3.14)" etc.
The numbered groups have the same meaning as above (except that $2 is now an empty string ("") when there is no leading sign, rather than null).
But be aware that it will also try to extract whatever numbers it can find. E.g., it will extract 127.0 from 127.0.0.1.
If you want something more sophisticated than that then I think you might want to look at lexical analysis instead of regular expressions. I'm guessing one could create a look-ahead-based expression that would recognize that "Pi is 3.14." contains a floating point number but Home is 127.0.0.1. does not, but it would be complex at best. If your pattern depends on the characters that come after it in non-trivial ways you're starting to venture outside of regular expressions' sweet-spot.
Paulpro and lbsweek answers led me to this:
re=/^[+-]?(?:\d*\.)?\d+$/;
>> /^[+-]?(?:\d*\.)?\d+$/
re.exec("1")
>> Array [ "1" ]
re.exec("1.5")
>> Array [ "1.5" ]
re.exec("-1")
>> Array [ "-1" ]
re.exec("-1.5")
>> Array [ "-1.5" ]
re.exec(".5")
>> Array [ ".5" ]
re.exec("")
>> null
re.exec("qsdq")
>> null
For anyone new:
I made a RegExp for the E scientific notation (without spaces).
const floatR = /^([+-]?(?:[0-9]+(?:\.[0-9]+)?|\.[0-9]+)(?:[eE][+-]?[0-9]+)?)$/;
let str = "-2.3E23";
let m = floatR.exec(str);
parseFloat(m[1]); //=> -2.3e+23
If you prefer to use Unicode numbers, you could replace all [0-9] by \d in the RegExp.
And possibly add the Unicode flag u at the end of the RegExp.
For a better understanding of the pattern see https://regexper.com/.
And for making RegExp, I can suggest https://regex101.com/.
EDIT: found another site for viewing RegExp in color: https://jex.im/regulex/.
EDIT 2: although op asks for RegExp specifically you can check a string in JS directly:
const isNum = (num)=>!Number.isNaN(Number(num));
isNum("123.12345678E+3");//=> true
isNum("80F");//=> false
converting the string to a number (or NaN) with Number()
then checking if it is NOT NaN with !Number.isNaN()
If you want it to work with e, use this expression:
[+-]?[0-9]+([.][0-9]+)?([eE][+-]?[0-9]+)?
Here is a JavaScript example:
var re = /^[+-]?[0-9]+([.][0-9]+)?([eE][+-]?[0-9]+)?$/;
console.log(re.test('1'));
console.log(re.test('1.5'));
console.log(re.test('-1'));
console.log(re.test('-1.5'));
console.log(re.test('1E-100'));
console.log(re.test('1E+100'));
console.log(re.test('.5'));
console.log(re.test('foo'));
Here is my js method , handling 0s at the head of string
1- ^0[0-9]+\.?[0-9]*$ : will find numbers starting with 0 and followed by numbers bigger than zero before the decimal seperator , mainly ".". I put this to distinguish strings containing numbers , for example, "0.111" from "01.111".
2- ([1-9]{1}[0-9]\.?[0-9]) : if there is string starting with 0 then the part which is bigger than 0 will be taken into account. parentheses are used here because I wanted to capture only parts conforming to regex.
3- ([0-9]\.?[0-9]): to capture only the decimal part of the string.
In Javascript , st.match(regex), will return array in which first element contains conformed part. I used this method in the input element's onChange event , by this if the user enters something that violates the regex than violating part is not shown in element's value at all but if there is a part that conforms to regex , then it stays in the element's value.
const floatRegexCheck = (st) => {
const regx1 = new RegExp("^0[0-9]+\\.?[0-9]*$"); // for finding numbers starting with 0
let regx2 = new RegExp("([1-9]{1}[0-9]*\\.?[0-9]*)"); //if regx1 matches then this will remove 0s at the head.
if (!st.match(regx1)) {
regx2 = new RegExp("([0-9]*\\.?[0-9]*)"); //if number does not contain 0 at the head of string then standard decimal formatting takes place
}
st = st.match(regx2);
if (st?.length > 0) {
st = st[0];
}
return st;
}
Here is a more rigorous answer
^[+-]?0(?![0-9]).[0-9]*(?![.])$|^[+-]?[1-9]{1}[0-9]*.[0-9]*$|^[+-]?.[0-9]+$
The following values will match (+- sign are also work)
.11234
0.1143424
11.21
1.
The following values will not match
00.1
1.0.00
12.2350.0.0.0.0.
.
....
How it works
The (?! regex) means NOT operation
let's break down the regex by | operator which is same as logical OR operator
^[+-]?0(?![0-9]).[0-9]*(?![.])$
This regex is to check the value starts from 0
First Check + and - sign with 0 or 1 time ^[+-]
Then check if it has leading zero 0
If it has,then the value next to it must not be zero because we don't want to see 00.123 (?![0-9])
Then check the dot exactly one time and check the fraction part with unlimited times of digits .[0-9]*
Last, if it has a dot follow by fraction part, we discard it.(?![.])$
Now see the second part
^[+-]?[1-9]{1}[0-9]*.[0-9]*$
^[+-]? same as above
If it starts from non zero, match the first digit exactly one time and unlimited time follow by it [1-9]{1}[0-9]* e.g. 12.3 , 1.2, 105.6
Match the dot one time and unlimited digit follow it .[0-9]*$
Now see the third part
^[+-]?.{1}[0-9]+$
This will check the value starts from . e.g. .12, .34565
^[+-]? same as above
Match dot one time and one or more digits follow by it .[0-9]+$

Javascript text manipulation

Just starting with js, decided to convert Friendfeed to a fluid app, and as part of that I need to be able to parse some numbers out of a string.
How do I complete this function?
function numMessages(text) {
MAGIC HAPPENS (POSSIBLY THE DARK ART OF THE REGEX)
return number;
}
input would be "Direct Messages (15)"
output would be 15.
Instincts tell me to find the first bracket then find the last bracket, and get the text in between but I don't know how to do that. Second instinct tells me to regex for [0-9], but I don't know how to run regexes in js. Jquery is avaliable already if needed.
Thanks!
This should do it:
>>> 'Direct Messages (15)'.match(/[0-9]+/g);
["15"]
Just be careful if you expect more than 1 number to be in the string:
>>> 'Direct 19 Messages (15)'.match(/[0-9]+/g);
["19", "15"]
If you only wanted the first match, you could remove the g flag:
>>> 'Direct 19 Messages (15)'.match(/[0-9]+/);
["19"]
If you only wanted to match what's between the parentheses
>>> 'Direct 19 Messages (15)'.match(/\((.*?)\)/);
["(15)","15"]
// first index will always be entire match, 2nd index will be captured match
As pointed out in the comments, to get the last match:
>>> var matches = 'Direct 19 Messages (15)'.match(/[0-9]+/g);
>>> matches[matches.length-1];
"15"
Though some boundary checking would also be appropriate. :)
var reg = new RegExp('[0-9]+');
reg.exec('Direct Messages (15)');
function numMessages(text) {
return text.match(/\d+/g);
}
This will return all numbers (\d is a special character class equivalent to [0-9]) from the string. the /g makes the regex engine do a global search, thereby returning an array of all matches; if you just want one, remove the /g. Regardless of if your expression is global or not, match returns an array, so you will need to use array notation to get at the element you want.
Note that results from a regular expression match are of type string; if you want numbers, you can use parseInt to convert "15" to 15.
Putting that all together, if you just want one number, as it seems to appear from your initial question text:
function numMessages(text) {
return parseInt(text.match(/\d+/)[0]);
}
str = "Direct Messages (15)";
numMessages(str); // 15

Categories