atof() question

Hi all,

I am getting the following warning using the following code:

Code:

---

float Convert(char *str)
{
    float num;
    num=atof(str);
    return num;
}

---

warning C4244: '=' : conversion from 'double ' to 'float ', possible loss of data

I thought that atof() is ascii to float..... not asci to double........
What am I doing wrong?

Thank you

>I thought that atof() is ascii to float.....
"float" being "floating-point", not the float type specifically. atof returns double.

Any suggestions to convert string (chars) to float type?

Code:

---

float Convert(char *str)
{
    float num;
    num = (float) atof(str);
    return num;
}

---

or

Code:

---

float Convert(char *str)
{
    float num;
    sscanf(str, "%f", &num);
    return num;
}

---

>Any suggestions to convert string (chars) to float type?
Why not just use double?

thank you for the code it's great!

I want to use float for study purpses.

Except the fact the float is 6 digit percision and double is 10,
I don't really understand the difference.

for example if I enter the following number 1234.56 in a float type the output will be 1234.560059 , I don't understand why.

Another thing if I use

Code:

---

num=(float)atof(string,"%f",&num);

---

In what point will I my data lost it's precision.

Many thanks again

Quote:

---

Originally Posted by salvadoravi

thank you for the code it's great!

I want to use float for study purpses.

Except the fact the float is 6 digit percision and double is 10,
I don't really understand the difference.

for example if I enter the following number 1234.56 in a float type the output will be 1234.560059 , I don't understand why.

Another thing if I use

Code:

---

num=(float)atof(string,"%f",&num);

---

In what point will I my data lost it's precision.

Many thanks again

---

http://docs.sun.com/source/806-3568/ncg_goldberg.html

Quote:

---

for example if I enter the following number 1234.56 in a float type the output will be 1234.560059 , I don't understand why.

---

A simple answer is "because computers work in base 2 and we (humans) work in base 10". Some numbers who are representable with a finite numbers of digits in base 10 aren't necessarily representable with a finite numbers of digits in base 2.

Example. "2/3" can't be represented with a finite number of digit in base 10. If we can take only 4 digits to represent that number, we'll write "0.667". But in base 3, you can write it as "0.2".

Quote:

---

Another thing if I use

Code:

---

num=(float)atof(string,"%f",&num);

---

In what point will I my data lost it's precision.

---

Doubles have more significative digits than floats. By the way, that call to atof is incorrect (mix between sscanf and atof).