JavaScript isNaN() Coerces Values When Testing For Numbers

Posted August 8, 2012 at 9:31 AM by Ben Nadel

Tags: Javascript / DHTML

Recently, I was tripped up in my JavaScript debugging due to the way that the isNaN() (ie. is "not a number") function treats its argument. I was trying to track down an elusive bug that would occasionally present a numeric ID as the empty string (""). To figure out why this was happening, I put an isNaN() check in place in order to perform logging. Except, the error kept happening and no logging was being performed.

After debugging my debugging code (ironic!), I discovered that calling isNaN() on the empty string actually returned "false"; that is, isNaN() considers the empty string to be a valid number. After reading up on the isNaN() documentation, I found out that isNaN() tries to coerce its argument into a number before returning a result.

Now, with coercion in mind, things started to make more sense:

  • Empty strings can be coerced into a falsey value.
  • Falsey values can be coerced into Zero.
  • Zero is a valid number.
  • Hence, isNaN() sees empty string as a number.

To see this in action, I put together a small demo that runs the empty string and some truthy / falsey values through isNaN(). I also played around with using parseFloat() as a means to bypass numeric coercion:

  • <!doctype html>
  • <html>
  • <head>
  • <meta charset="utf-8" />
  • <title>Testing JavaScript isNaN() Value Coercion</title>
  • <script type="text/javascript">
  •  
  •  
  • // Empty strings can be coerced to falsey values. And all
  • // truthy / falsey values can be coerced to numbers. As such,
  • // truthy / falsey values are *not* not numbers.
  •  
  • console.log( "Empty String isNaN: ", isNaN( "" ) );
  •  
  • console.log( "False isNaN: ", isNaN( false ) );
  •  
  • console.log( "True isNaN: ", isNaN( true ) );
  •  
  •  
  • // -------------------------------------------------- //
  • // -------------------------------------------------- //
  • console.log( ". . . . ." );
  • // -------------------------------------------------- //
  • // -------------------------------------------------- //
  •  
  •  
  • // ParseFloat() coerces arguments to a String value and then
  • // tries to extract the numeric value. As such, it doesn't
  • // fall victim to truthy / falsey coercion.
  •  
  • console.log( "Empty String isNaN: ", isNaN( parseFloat( "" ) ) );
  •  
  • console.log( "False isNaN: ", isNaN( parseFloat( false ) ) );
  •  
  • console.log( "True isNaN: ", isNaN( parseFloat( true ) ) );
  •  
  •  
  • // -------------------------------------------------- //
  • // -------------------------------------------------- //
  • console.log( ". . . . ." );
  • // -------------------------------------------------- //
  • // -------------------------------------------------- //
  •  
  •  
  • // Simply exploring how parseFloat() coerces values (forcing
  • // them to be Strings before it parses them).
  •  
  • var complexObject = {};
  •  
  • // Override the core to-string value.
  • complexObject.toString = function(){
  •  
  • // Return a NUMBER string.
  • return( "012345" );
  •  
  • };
  •  
  • // Test float parsing.
  • console.log(
  • "Complex Object Float: ",
  • parseFloat( complexObject )
  • );
  •  
  •  
  • </script>
  • </head>
  • <body>
  • <!-- Left inentionally blank. -->
  • </body>
  • </html>

When we run the above code, we get the following console log output:

Empty String isNaN: false
False isNaN: false
True isNaN: false
. . . . .
Empty String isNaN: true
False isNaN: true
True isNaN: true
. . . . .
Complex Object Float: 12345

As you can see, all truthy / falsey values are seen as "numbers" by the isNaN() function. Furthermore, trying to use parseFloat() as an intermediary cuts out the numeric type coercion because parseFloat() coerces its values into Strings before parsing them.

I guess this is the documented behavior of isNaN(); but, seeing as I've used the isNaN() function maybe 3 times in my career, I wasn't fully aware of how it worked. But now I know - and knowing is half the battle!



Reader Comments

There are no comments posted for this web log entry.

Post A Comment

Comment Etiquette: Please do not post spam. Please keep the comments on-topic. Please do not post unrelated questions or large chunks of code. And, above all, please be nice to each other - we're trying to have a good conversation here.

Please review the following issues:

Author Name:


Author Email:

Author Website:

Comment:

Supported HTML tags for formatting: <strong>bold</strong>   <em>italic</em>   <code>code</code>







  • Help Wanted - Find Your Next ColdFusion Job
Ben Nadel's Company - Epicenter Consulting Recent Blog Comments
May 25, 2013 at 10:08 AM
Using "//" And ".//" Expressions In XPath XML Search Directives In ColdFusion
@Ben, my question is that i want the current node with its tag and its parent node. i just want only that data. So, give me the solution for that. and remember solution is working on " xpath 1.0 ... read »
May 25, 2013 at 10:01 AM
Using "//" And ".//" Expressions In XPath XML Search Directives In ColdFusion
hey ben, i want get my current node tag and also want the root node tag withing. So, how can i fix it.. ! ... read »
May 24, 2013 at 5:39 PM
Ask Ben: Manually Enforcing Basic HTTP Authorization In ColdFusion
@Adam Oops! My mistake! I hadn't gotten that far in my testing - I'm still baby stepping my way through the process. ... read »
May 24, 2013 at 5:13 PM
Ask Ben: Manually Enforcing Basic HTTP Authorization In ColdFusion
Hi Jason, Thanks for checking up on that, but I still stand firm on my position. :) There are actually two listLast()'s in use, and you're right that the one using a space as a delimiter is fine. ... read »
May 24, 2013 at 4:45 PM
Ask Ben: Manually Enforcing Basic HTTP Authorization In ColdFusion
@Ben I have been lurking your site for quite some time, and haven't stepped up to comment until today. Thanks for all the great info - keep it up! @Adam I believe you are mistaken... as the commen ... read »
May 24, 2013 at 11:21 AM
Strange Interaction Between DeserializeJson(), ArrayContains(), And Database Values In ColdFusion
@WebManWalking, Ha ha, let's us never speak of justifying "##" notation again :P ... read »
May 24, 2013 at 11:18 AM
Strange Interaction Between DeserializeJson(), ArrayContains(), And Database Values In ColdFusion
@Ben, Ah, so it was indeed how I vaguely remembered it to be: A direct assignment value = users.id[ i ] causes value to retain the sticky datatype of the query column. Although unnecessary in ... read »
May 24, 2013 at 9:11 AM
Preventing Links In Standalone iPhone Applications From Opening In Mobile Safari
@Brandon, Hi, No, I haven't been able to do that. I have just kept it as it is. ... read »
InVision App - Prototyping Made Beautiful With Prototyping Tools