Thread: 5 Byte Integer

I have a requirement from a customer to create a binary data file, with the first 5 Bytes (or 40 bits) representing the microseconds of the day. Since there are 86400 seconds per day, the max microseconds of the day is 86400000000 and the customer wants this number represented in 5 bytes. How can I store this number into a 5 byte integer? Longs only go up to 4 bytes, or 32 bits. I could create an array of 5 unsigned chars, but how to store the number?

The number is read as a double, and would look something like 742039.302502 and I would somehow get it into a 5 byte integer representation of 742039302502

Any ideas?

thanks,

NewPappa

1) Microsoft Visual C/C++... on Windows platform.

2) The microseconds is read from a binary file that has data that has been recorded already. The customer wants the data in this binary file processed and outputted to a binary file with the 5 byte microseconds time tag at the beginning of each write to file.

Any ideas on how to solve this problem?

thanks.

it's stored in the first file as a double, then must be converted and outputted to the 2nd file in it's 5 byte integer representation.

Originally Posted by erikj

Maybe this will work

Code:

char *p;
char buf[32];
double dbl;

fread( dbl, 1, sizeof(double), fp );

sprintf(buf, "%f\0", dbl);

if((p= strchr(buf, '.'))!=NULL)
    memcpy(p, p + 1, strlen(p) + 1);
	 
for(p = buf + strlen(buf); *p == '0'; p--)
   *p='\0';

fwrite( buf, 1, strlen(buf), out );

That just stores it as an ascii value, I need to store it in it's integer representation. For example a value of 29 would look like a 0x1D in a binary file. Now how to represent a large number like 81556123456, which needs more than 32 bits of precision??

Thanks.

I finally came up with a solution. It required using the __int64 data type. I just didn't know about this data type. Having that data type at my disposal made the solution a piece of cake.

thanks to all that helped.

NewPappa

>> Except that is not 40 bits, it's 64. Thus, if you write it to a file, you'll be writing a binary 64 bits, instead of 40, which when you read it again, will screw it all up.

But I can create a union with an 8 element unsigned char array, and use the 5 bytes of significance.

Code:

union mixed {
  unsigned char byteArray[8];
  __int64 iPortion64;
};

union mixed microSeconds;

something like that.