Altering decimal places of float values read from Binary file

I am reading a binary file containing records of information. Part of the record is a Account Balance which is a float value. It is already stored with the default 6 decimal places.

Does anyone know how I can alter the stored value to only be 2 decimal places. As I am trying to total up all the balances, and the 6 decimals places is altering my final figure.

This is the line I have that reads the balance from the binary file, but from this point it is stored in the structure member as a 6 decimal place float value.

Code:

---

/*
fread(&customer.cust_bal, sizeof(customer.cust_bal), 1, custmast_ptr); */

---

Any ideas??
Kelly

round the float value to 2 decimals before use :

Code:

---

#include <math.h>
float round(float f)
{
 return floor(f+0.5);
}

...

float amount = 1.23456;
amount = round(amount*100)/100;
printf ("1.23456 rounded to 2 decimals is : %f", amount);

...

---

I think your are having trouble formatting your output. So try:

Code:

---

printf(" %3.2f\n", f);    /*  f is a float variable  */

---

And yes, your fread() has got nothing to do with your 'final figure'. Even if you store the value 1.200000 as 1.2 in your variable, , internally, it wil still be stored as 1.200000

no, it wasnt the output that was the problem, it was when the data was read from the binary file into the float variable.

Thanks to everyone, I seem to have (fingers crossed) sorted it by redefining the FLT_DIG constant changing the default number of decimal places floats are stored from 6 to 2.

seems to be OK so far

Kelly :-)

Be very careful when you modify the standard header files for your compiler. Bad things(tm) could happen.

-Prelude

Thanks

It turned out that actually that hadnt sorted out the problem!

so im back to square one!

Basically for the first record it is reading in -2407.629,
I want to STORE the value as -2407.63

Anyone have any idea of any functions that will round up/ round down as necessary?

> It is already stored with the default 6 decimal places
No it isn't
You've written out the 32 bits which comprise the floating point number, and these have a precision of 6 decimal digits.
This means that
0.123456 is accurately stored
123.456 is accurately stored
but
123.456789 is rather more random in the last 3 decimal digits.

There is nothing you can do to a float which would make it behave as you would want.

If you really want 2 decimal places, then I suggest you multiply your float by 100, cast the result to an integer, and perform the summation using integers. At the end, divide the result by 100 to get your sum back into floating point format.

But if all you want to do is print it, try
printf( "%d.%02d", sum/100, sum % 100 );
Then you won't have any random noise from the least significant bits of your float.

Hi, I think I know the program your working on...424project by any chance...sent pm.. mail me if you require help.

well I've tried conversion as well now, and that still doesnt work. Perhaps I can better illustrate this way. :mad:

Here is the first 5 float values, and the increasing total balance as they are added.

-2407.62915 -> -2407.62915
99173.45313 -> 96765.82398
988.530029 -> 97754.354
700 -> 98454.354
-4729.879883 -> 93724.47412

However when I add the float values I am getting different total values from the first addition ( done simply by total += value);

-2407.62915 -> -2407.62915
99173.45313 -> 96765.82031
988.530029 -> 97754.35156
700 -> 98454.35156
-4729.879883 -> 93724.46875

Hope I've explained the problem a little better this time. Please please please, any ideas on what the problem is !
:confused:

Are you by any chance trying to accumulate the total customer balance of all customers on the new cust mast file in prog 4 424C.

have just pm'd you bigtamscot. Email me directly if you can help at all to the email address in the message

THANK YOU!