Rounding and truncating numbers in JavaScript
Rounding and truncating is a bread and butter action for every single developer. It was covered during your first few math lessons in primary school. Hopefully you still remember how it works in the world of numbers.
5 or more? Raise the Score. 4 or less? Let it Rest.
Let’s use this knowledge and translate it to JavaScript using the built-in object called Math. As the name can suggests, it has a collection of properties and methods for mathematical operations on numbers. There is one small difference between Math and other built-in global objects. Math isn’t a constructor which means that all properties and methods that belong to it are static (meaning that they need to be called by using Math as an object).
Rounding vs Truncating
The difference between these two methods is minor but very important to understand. Both of them are methods of approximating a number by dropping decimal places. Rounding approximates a number using a nearby number at a given degree of accuracy. It can occur in two directions: up and down. Rounding up approximates a number towards positive infinity. Rounding down towards negative infinity. Truncating approximates without rounding. In other words, it “rounds” towards zero.
Rounding
3.14159 ≈ 3.1416
Truncating
3.14159 ≈ 3.1415
Hopefully you get the difference. It makes truncating rarely useful in precise calculations (although JavaScript probably isn’t a good choice at all if you need precise calculations) but you can come across a situation when it may be irreplaceable. Once example can be when needing to drop decimal places from a pixel value to avoid anti aliasing or weird pixel rounding which is completely different across browser engines.
Rounding numbers in Javascript
Rounding is straight forward. We can round to the nearest integer, round down or round up. JavaScript uses three methods to achieve this:
Math.round()- rounds to the nearest integer (if the fraction is 0.5 or greater - rounds up)Math.floor()- rounds downMath.ceil()- rounds up
Math.round(3.14159) // 3
Math.round(3.5) // 4
Math.floor(3.8) // 3
Math.ceil(3.2) // 4
Rounding numbers with decimal precision requires a little bit of calculation and Math.round(). Optionally we can use the toFixed() method that belongs to the Number prototype. The output type of toFixed() is a string which needs to be passed to a top-level function called parseFloat() to return a number. Unfortunately this seems to be really slow.
Math.round(3.14159 * 100) / 100 // 3.14
3.14159.toFixed(2); // 3.14 returns a string
parseFloat(3.14159.toFixed(2)); // 3.14 returns a number
Truncating numbers in Javascript
Math.trunc() simply removes all the fractional digits. It takes one argument which is a number. If the argument is a positive number it behaves exactly the same as Math.floor(). For negative numbers it does the same job as Math.ceil().
Math.trunc(3.14159); // 3
Math.trunc(-3.14159); // -3
It’s worth mentioning that the browser support for Math.trunc() isn’t great. It is part of new ES2015 (yeah, I prefer ES6 too) specification. Have a look at the browser support list:
- Google Chrome >= 38
- Firefox >= 25
- Internet Explorer >= Nope :(
- Opera >= 25
- Safari >= 7.1
Luckily there is a way to use this without ES6 support (thanks to Johny who suggested this solution in comments below). We can use bitwise operators to accomplish this task. Unfortunately there are some restriction as well. All bitwise operations work on signed 32-bit integers. Using them converts a float to an integer. In practice it means that we can safely work up to 2^31−1 (2 147 483 647) which is much less than Number.MAX_VALUE (1.7976931348623157e+308). This isn’t a great idea for monetary calculations either.
3.14159 | 0; // 3
-3.14159 | 0; // -3
3000000000.1 | 0 // -1294967296 Ups :(
TLTR (too long to read)
I know, I know - time is money. Lets sum it up.
Math.round()- rounds to the nearest integerMath.floor()- rounds down towards negative infinityMath.ceil()- rounds up towards positive infinityMath.trunc()- rounds up or down towards zero (bad browsers support)
Widely-supported way of truncating is bitwise operator:
-3.99 | 0
Although what's done here is not entirely clear. (bitwise shift operates on 32-bit numbers only so it returns those [integers])
You can write your own truncating polyfill using just that.
function truncate(number){ return number|0; }
See which is faster for yourself here. :)
http://jsperf.com/bitwise-v...
Fantastic tip. I may update this article with your tip if you don't mind. I won't forget about kudos for Johny :) Thanks again, really nice hint!
No problem. :)
Thanks a lot again. I just updated article and applied all your suggestions.
https://github.com/pawelgrz...
Your demo link is broken but my testing showed this
|0method to slightly faster than the rest (and the shortest to type).Math.trunc,Math.roundetc are all about equal.toFixed(0)is exponentially slower (even before converting back to a number) and it also rounds kinda weird in some cases, like how(0.35).toFixed(1) = 0.3yet(1.35).toFixed(1) = 1.4I will fix the link. Thanks for pointing this one out and sharing your results!
Another thing that came to my mind is that toFixed is returning a string, not a number. So to get the number you would have to use parseInt (slow), or another method of getting fixed numbers (fast):
Math.round( 3.66666 * 100 ) / 100
Check the speeds: http://jsperf.com/tofixed-p...
Yeah I know what do you mean. I think you wanted to mention parseFloat() not parseInt() because in this example is all about floating points :-) Worth to add to article - definitely. Thank you so much for your input Johny. Can you give me your Twitter handle (if you have one). I'll update this article with all your advices later on today.
Rounding and truncating are really necessary for a developer to know and it's a good guide that you have created this article which can guide them on how to do it in Javascript.
Thank you very much! I'm glad that you found it useful.
toFixed is unreliable for rounding. Try this:
1.265.toFixed( 2 )
parseFloat( 1.265.toFixed( 2 ) )
// round to the nearest number taking into account the sign
const round = number => Math.sign(number) * Math.round(Math.abs(number))
// rounds a number to its closer taking into account sign
const round = number => {
if (number === 0) return 0;
return number > 0
? Math.round(number)
: Math.sign(number) * Math.round(Math.abs(number));
}
That was a presentation!
Hi I have a problem with JavaScript
I have currency value 11.26 --I need always roundup ---12
If I have 12.56 I need roundup 13.
Thanks in advance.
Hi.
This is exactly what Math.ceil() is doing.
Thanks
Someone didn't read a word of this post :P
The
Math.round(3.14159 * 100) / 100 // 3.14rounding approach will not always be accurate. There are a lot of situations where it will round down values:Math.round(Math.round(0.145 * 1000) / 10) / 100 // 0.15
Math.round(Math.round(1.005 * 1000) / 10) / 100 // 1.01
Hi,
How can I change 9.26 to 9.27 using math.round ?
I am not sure if you understand the concept of rounding. It is not made for changing values but for rounding values.
`9.26 + 0.01`
Math.round rounds it to 9.
If you want 9.27 then,
You can do this by using parse float Method.
Maybe:
* parseFloat(9.264.toFixed(2)); //9.26
* parseFloat(9.266.toFixed(2)); //9.27
var truncated = number - number % 1;var rounded = number + 6755399441055744.0 - 6755399441055744.0; // 2^51 + 2^52This works for numbers with absolute value < 2 ^ 51.
var original=224.985
document.getElementById("demo").innerHTML = Math.round(original*100)/100;
var original2=224.98499999
document.getElementById("demo2").innerHTML = Math.round(original2*100)/100;
output
224.99
224.98
in second case when no is 999999, it is not rounding to upper one
What to be done to get second operation's value same as first one?
https://t.co/sO86kEe3uX?Vqkugn
https://t.co/yc6wKCq1xP?cjg...
Great article , thanks ! I wanted to ask though about this issue I encountered : somehow in Javascript : '28.99 - 28.70' is equal to => '0.28999999999999915' and not 0.29 ! How do I approach this to always get 0.29 ? 0.28999999999999915
Hi Klaud.
You just became a victim of the "floating-point arithmetic precision". For people, something as obvious may not be that obvious for computers because of how they deal with number. Google this term to find out more about it. A classic example!
How to deal with it then? The answer isn't that simple, but it all comes down to avoid dealing with fractions if you can. In many e-commerce solutions, prices are represented like this…
By operating with numeric values using notations like this, all operations are done on the
centAmmountfirst, and then the final result is divided by thefractionDigits. Hopefully, that helps.Nice This is and i have used this in my javascript code
I am glad it helped you out :)