problem facing with modf function in skill

This is because of floating point rounding errors. It's nothing to do with SKILL (other languages will have similar problems) - similar questions (maybe not with modf but with other calculations) have come  up on numerous occasions on the forums.

You should never do an equal type check on floating point numbers because it's highly likely they are not precisely equal. Floating point numbers are represented in IEEE floating point format which uses a binary mantissa and so .1 in decimal ends up being a recurring fraction (similar to how 1/3 is 0.3333 recurring in decimal). Because of this, you easily can end up with a rounding error due to a loss of a bit in the 53rd binary digit of the mantissa.

To illustrate this:

modf(0.945 0.315)-0.315
-5.551115e-17

So as you can see, it's very slightly less than 0.315 but when displayed with fewer digits of precision, it appears to be 0.315. If I generate the 0.945 slightly differently:

modf(0.315*3 0.315)
5.551115e-17

Even now you see it's not quite 0 - slightly above. Another example:

0.315*3-0.945
1.110223e-16

What  you have to do is do equality checks using a range of values - sometimes things like:

abs(answer - expectedAnswer) < someTolerance

For your modf, if the numbers have three digits of precision, maybe you'd be better multiplying by 1000, rounding, and then using mod() instead (i.e. integer modulus)/

Regards,

Andrew.

Thanks alot. I did not thought like that. That will help me.

Thanks alot. I did not thought like that. That will help me.