How does one enumerate through all doubles?

Question

4.00/5 (1 vote)

See more:

I have some code and I want to verify that it works for all doubles. Since I am not the trusting sort I would like to have it run through all the doubles and test if my method works. I'm converting the double to another datatype and then converting it back.

Unfortunately this is not a trivial task like going through Int32. If anyone has done this or knows where some code is that does this please let me know.

Thanks
Ken

Posted 10-Jun-10 6:44am

KenJohnson

Add a Solution

2 solutions

Solution 1

All you have to do is test with three or so values. Best case, worst case, and a couple of randomly chosen middle-of-the-road values.

Besides, when you say enumerate through all the doubles, do you mean every possible value? That test will take quite a while to run.

Posted 10-Jun-10 6:47am

#realJSOP

Comments

KenJohnson 10-Jun-10 13:44pm

Thanks for responding. I am converting the doubles to a significantly more fine grained numeric type. I want to make sure that one can do a round trip with no loss of information so I would like to test every one. I am aware that the running time of the test may be considerable, but I have an extra computer that I would be willing to let churn for as weeks if necessary.

#realJSOP 10-Jun-10 14:29pm

If you want accuracy, you should use decimal instead of double. decimal values take a bit longer to work with, but if accuracy is your goal, that's the way to go. With doubles, the minute you do math on them, you can't be entirely sure of what the result will be (1.0 x 1 may result in 1.00000000001).

Add a Solution

Add your solution here

Treat my content as plain text, not as HTML

Preview 0

…

Existing Members

Sign in to your account

...or Join us

Download, Vote, Comment, Publish.

Your Email
Password
Forgot your password?

Your Email
This email is in use. Do you need your password?
Optional Password

I have read and agree to the Terms of Service and Privacy Policy
Please subscribe me to the CodeProject newsletters

When answering a question please:

Read the question carefully.
Understand that English isn't everyone's first language so be lenient of bad spelling and grammar.
If a question is poorly phrased then either ask for clarification, ignore it, or edit the question and fix the problem. Insults are not welcome.
Don't tell someone to read the manual. Chances are they have and don't get it. Provide an answer or move on to the next question.

Let's work to help developers, not make them feel stupid.

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)

Andrew Rissing · Accepted Answer · 2010-06-10T08:09:00

I'd recommend incrementing a long and then converting it to a double, using the below code:

C#

double dbl;
 
for (long lng = long.MinValue; lng < long.MaxValue; ++lng)
{
    dbl = BitConverter.Int64BitsToDouble(lng);
 
    /* Test the double here */
}
 
// Also test for long.MaxValue as well.
dbl = BitConverter.Int64BitsToDouble(long.MaxValue);

It'll keep you from potentially underflowing if you were to just take a double and increment it by some number.

Be aware, you'll get some special 'case' values from doing this, like infinity and NaN and such. You'll probably want to use the static methods on double to remove those cases or just constrain the value between Min/Max. Something like that...

Have fun with your ~18,446,744,073,709,551,616 test cases. :D