Why does parenthesis around a fraction (5/9) break my program? Beginner!

Hey, I just started trying to learn programming so please be gentle. I tried to think of some easy practice programs to make and conversion came to mind so I tried to make a little temperature conversion program. In the conversion equation there's a fraction 5/9 and I put parenthesis around it because I didn't think it would change the result and I thought it made it easier to read. The parenthesis make the result always come out to be zero. When I remove the parenthesis the result is correct. Why is this?

Code:

---

#include <stdio.h>
#include <stdlib.h>


int main()
{
    float fDegrees;
    float cDegrees;
    char fDegreesString[20];
    char cDegreesString[20];
    char desiredConversion;


    printf("Would you like to convert from Celcius or Farenheit? \nEnter C or F.\n");
    desiredConversion = getchar();
    getchar();


    if (desiredConversion == 'C' || desiredConversion == 'c')
    {
        printf("What is the temperature in C?\n");
        gets(cDegreesString);
        cDegrees = atoi(cDegreesString);
        fDegrees = (cDegrees * (9 / 5)) + 32;
        printf("The temperature in F is %.1f degrees.\n", fDegrees);
    }


    else if (desiredConversion == 'F' || desiredConversion == 'f')
    {
        printf("What is the temperature in F?\n");
        gets(fDegreesString);
        fDegrees = atoi(fDegreesString);
        cDegrees = (fDegrees-32)*(5/9);
        printf("The temperature in C is %.1f degrees.\n", cDegrees);
    }


    else
        printf("You must type C or F.\n");


    return 0;
}

---

Any other suggestions you have to make this better would be great. GCC shows a warning that I shouldn't use gets. Any suggestion on what I should use instead?

Thanks!

I reckon (5/9) gets truncated into an integer, whereas without the parentheses the compiler sees the context of the whole statement and doesn't do so.

Code:

---

float i = 5 / 9; // 0.555
float j = (5 / 9); // (int / int) == int(0.555) == 0

---

Multiplying by that 0 will thus always give you 0.

By making one of the operands a floating-point literal, you can correct this:

Code:

---

float k = (5.0 / 9);

---

(5/9) results in an integer value, NOT a floating point value. You want ((double)5/9). Cast 5 to a double, then 9 will also be promoted, resulting in a double result.

gets() has been depreciated in C99, and eliminated in C11. You should NEVER use gets(), but use fgets() instead.

To elaborate, it's because of order of operations, and the fact that, when an expression has mixed integers and floating-point operands, the integers are promoted into floats to do the math.

Code:

---

(fDegrees-32)*5/9;

---

In this case, fDegrees is a float, so 32 is converted to 32.0 and the result is a float. Next, the float is multiplied by 5, so 5 is promoted to a float, too. Finally, 9 is promoted to a float as floating-point division is done.

Code:

---

(fDegrees-32)*(5/9);

---

Here fDegrees-32 is still a float, but the parentheses mean that 5/9 is evaluated before the multiplication. Since 5 and 9 are integer literals, it does integer division, and 5/9 = 0 in integer division.

The simplest way to fix this is to use floating-point literals such as (5.0f/9.0f)

You also likely want atof() not atoi() as you're trying to parse floating-point values, not integers.

This is great information! I'm familiar with the concept of int/float/double/decimal but I hadn't thought of needing to specify the granularity of literal numbers. Typecasting is a concept I'm aware of but I didn't realize that it was something I could/should do with literal numbers. I only thought of it pertaining to variables.

I played around with different ways of diving numbers to see what different results I get. It appears that all numbers are assumed to be integers unless you specifically state them to be something else. I'm sure that's probably common knowledge but I had't learned that yet.

Code:

---

#include <stdio.h>


int main()
{
    float a = 7;
    float b = 11;
    float c = 0;


    c = 7 / 11;// = 0
    c = a / b;// = 0.636364
    c = (float) 7 / (float) 11;// = 0.636364
    c = 7.0f / 11.0f;// = 0.636364


    return 0;
}

---

Quote:

---

It appears that all numbers are assumed to be integers unless you specifically state them to be something else.

---

5 is an integer, since no decimal point appears in the constant.
5.0 or 5. is a double since the decimal point is part of the constant.

There is no "assumption" about it.

You might want to review the chapters on all forms of data.

Quote:

---

Originally Posted by Cat

To elaborate, it's because of order of operations, and the fact that, when an expression has mixed integers and floating-point operands, the integers are promoted into floats to do the math.

Code:

---

(fDegrees-32)*5/9;

---

In this case, fDegrees is a float, so 32 is converted to 32.0 and the result is a float. Next, the float is multiplied by 5, so 5 is promoted to a float, too. Finally, 9 is promoted to a float as floating-point division is done.

Code:

---

(fDegrees-32)*(5/9);

---

Here fDegrees-32 is still a float, but the parentheses mean that 5/9 is evaluated before the multiplication. Since 5 and 9 are integer literals, it does integer division, and 5/9 = 0 in integer division.

The simplest way to fix this is to use floating-point literals such as (5.0f/9.0f)

You also likely want atof() not atoi() as you're trying to parse floating-point values, not integers.

---

I was going to point out he was using the wrong formula and that he should also be using atof but you beat me to it. For the OP though, here are some other formulas you can use as well.

C = (F - 32) * 5 / 9 (this is the one you're trying to use)
C = (F - 32) / (9 / 5) (notice how this one requires the parenthesis around the 9/5. The division of 9/5 must get divided into whatever F-32 is. We don't want F-32 getting divided by just 9).
C = (F - 32) / 1.8

Thanks for all of the information, this is great!

To sum it all up:
-I should be using atof instead of atoi. This totally makes sense. I'm dealing with floats throughout the whole thing, it's silly not to use atof.
-I should use fgets instead of gets. fgets grabs the \n from hitting enter in the console. I probably have to do a little shenanigans to get rid of that if I don't want to use the \n.
-I should use doubles or floats by putting 5.0 instead of 5 or typecasting the integers as floats. I definitely will look more into basic forms of data

The only thing I don't get is how is the formula I'm using wrong? I wrote this up and I get the same answer using all four versions of the formula; 37.777779

Code:

---

#include <stdio.h>


int main()
{
    float c1;
    float c2;
    float c3;
    float c4;


    float f = 100;


    c1 = (f - 32.0) * 5.0 / 9.0;
    printf("%f\n", c1);
    c2 = (f - 32.0) * (5.0 / 9.0);
    printf("%f\n", c2);
    c3 = (f - 32.0) / (9.0 / 5.0);
    printf("%f\n", c3);
    c4 = (f - 32.0) / 1.8;
    printf("%f\n", c4);


    return 0;
}

---

Quote:

---

Originally Posted by sandwichface

Thanks for all of the information, this is great!

To sum it all up:
-I should be using atof instead of atoi. This totally makes sense. I'm dealing with floats throughout the whole thing, it's silly not to use atof.
-I should use fgets instead of gets. fgets grabs the \n from hitting enter in the console. I probably have to do a little shenanigans to get rid of that if I don't want to use the \n.
-I should use doubles or floats by putting 5.0 instead of 5 or typecasting the integers as floats. I definitely will look more into basic forms of data

The only thing I don't get is how is the formula I'm using wrong? I wrote this up and I get the same answer using all four versions of the formula; 37.777779

Code:

---

#include <stdio.h>


int main()
{
    float c1;
    float c2;
    float c3;
    float c4;


    float f = 100;


    c1 = (f - 32.0) * 5.0 / 9.0;
    printf("%f\n", c1);
    c2 = (f - 32.0) * (5.0 / 9.0);
    printf("%f\n", c2);
    c3 = (f - 32.0) / (9.0 / 5.0);
    printf("%f\n", c3);
    c4 = (f - 32.0) / 1.8;
    printf("%f\n", c4);


    return 0;
}

---

---

I might of worded that a bit better. When I said you was using the wrong formula, what I meant was you're using:

Code:

---

cDegrees = (fDegrees-32)*(5/9);

---

As Cat has previously noted, one of your problems lie in your parenthesis at the end there.

Code:

---

cDegrees = (fDegrees-32)*(5/9);

---

and

Code:

---

cDegrees = (fDegrees-32)*5/9;

---

are two different formulas. On my system, if you remove the parenthesis at the end there like:

Code:

---

cDegrees = (fDegrees-32)*5/9;

---

it works fine.

Oh, the reason the second example works is because you're doing the type cast to float by putting a .0 after the decimal. You're no longer dividing an integer by an integer. You're now dividing a float by a float. Previously, by removing the parenthesis, you get this:
cDegress = (fDegrees-32) * 5 / 9;

We do parenthesis first
float = (float - int) * int / int;

(float - int) got promoted (converted) to a float and now we're multiplying it the int data type
float = (float * int) / int;

(float * int) got promoted to a float and now we're doing the division by an int data type
float = float / int;

int got promoted to a float when we divided so the answer is now a float data type
float = float;

The way the C compiler handles operations like this is because of the size of the different variables. Int is smaller than float. So when you do something like float - int, int gets converted into float because the int datatype might not be able to hold the variable length or you would loose the precision (the stuff after the decimal point). Anything in parenthesis gets handled first. With your first way, it would of looked something like this:

cDegress = (fDegrees-32) * (5 / 9);

First the compiler processes (float - int) and then it processes (int / int)
float = (float - int) * (int / int);

Because with the int data type, we must have whole numbers, so (int / int) becomes 0
float = (float) * 0

anything multiplied by 0 is 0.
float = 0


Also, C = (F - 32) / (9 / 5) is the formula, but in C, you would need for this to work successfully, C = (F - 32) / (9.0f / 5) or C = (F - 32) / (9 / 5.0f) or C = (F - 32) / (9.0f / 5.0f)

I hope this clears things up a bit and I'm sorry for the confusion.

Quote:

---

Originally Posted by sandwichface

-I should use fgets instead of gets. fgets grabs the \n from hitting enter in the console. I probably have to do a little shenanigans to get rid of that if I don't want to use the \n.

---

I usually define a function called chomp that gets rid of a trailing new line if it exists. Here is how you define it:

Code:

---

#include <string.h>

void chomp(char *buf)
{
    char *p = strrchr(buf, '\n');
    if (p) *p = '\0';
}

---

Use it like this:

Code:

---

#include <stdio.h>

#define MAXCOUNT 1000
char linebuf[MAXCOUNT];

int main(void)
{
    while (fgets(linebuf, MAXCOUNT, stdin) != NULL) {
        chomp(linebuf);
        printf("You said \"%s\".\n", linebuf);
    }
    return 0;
}

---