-
October 13th, 2003, 07:45 AM
#1
double and %
Hi,
I have a very big number (up to 15 digits) which is stored as a double
I need to know if this number is integer or not
but % function doesn't exist for double
I wanted to do chaeck myDoubleNum%1 is not equal to 0
any suggestions?
thanks
-
October 13th, 2003, 08:41 AM
#2
Given that doubles cannot represent all values with absolute precision you might find that you assign (or enter) an integer (say 15.0) but that the double represents it as 14.999999999999.
What are you trying to do?
Dave Mclelland.
-
October 13th, 2003, 08:49 AM
#3
There is a possible solution, but it has a serious caveat.
The solution is to use the function modf
modf
Splits a floating-point value into fractional and integer parts.
double modf( double x, double *intptr );
Return Value
This function returns the signed fractional portion of x. There is no error return.
The problem with this is that the fractional part may not be exactly zero, if you have worked with the number a bit. This is a standard problem when handling floating point numbers and the usual solution is to treat everything smaller than a given epsilon as zero.
So the code could look like this:
Code:
#include <math.h>
const double epsilon = 0.00001; // you should adjust this as needed
bool IsInteger(double d)
{
double intpart;
return fabs(modf(d, &intpart)) < epsilon;
}
Get this small utility to do basic syntax highlighting in vBulletin forums (like Codeguru) easily.
Supports C++ and VB out of the box, but can be configured for other languages.
-
October 13th, 2003, 08:51 AM
#4
well
I found an ugly soloution to my specific problem
I get my gig number as a string and I have to divide it by 100 before using it.
since the string doesn't contain . (that's one of the rules)
then I simply check its last 2 chars I need them both to be 00 inorder to get an Integer
something like that:
double dMyDouble = atof(sMyTmpString.c_str());
dMyDouble /= 100;
if((sMyTmpString[14]== '0') && (sMyTmpString[13]== '0'))
{
//an integer
}
else
//not an integer
note: sMyTmpString is maximum 15 chars long and does not '.'
What do you think?
-
October 13th, 2003, 08:57 AM
#5
Why don't you just use an integer in the first place ? Many compilers include 64 bit integers which would be enough to handle 15 decimal digits. On MS VC, the datatype in question is called __int64.
Get this small utility to do basic syntax highlighting in vBulletin forums (like Codeguru) easily.
Supports C++ and VB out of the box, but can be configured for other languages.
-
October 13th, 2003, 09:32 AM
#6
well I can't use int because the number may not be an Integer and then I need to treat it differently, but you gave ma another idea, check the difference between the double value to the __int64 value, meaning:
double dMyDouble = atof(sMyTmpString.c_str());
dMyDouble /= 100;
double dDiff = dMyDouble - (__int64)dMyDouble;
if(dDiff)
printf("Not an integer");
else
printf("An integer");
What do you think?
Is it better than my privious one?
-
October 13th, 2003, 03:01 PM
#7
There may be problems with both approaches. The first one will not work correctly if the string doesn't contain exactly 14 digits. What if the string is only "12" ? The second approach suffers from the same rounding problems that Dave and myself have pointed out above.
If you really need doubles then just use my IsInteger function.
Get this small utility to do basic syntax highlighting in vBulletin forums (like Codeguru) easily.
Supports C++ and VB out of the box, but can be configured for other languages.
-
October 13th, 2003, 04:49 PM
#8
what about:
double d = 5.1;
if( double(int(d)) == d ) // is a integer
-
October 13th, 2003, 06:06 PM
#9
Originally posted by mwilliamson
what about :
double d = 5.1;
if( double(int(d)) == d ) // is a integer
Same representation problem. Floating point numbers may not exactly correspond with what you assign them. You cant safely do an == with a double, its not a precise representation.
Dave Mclelland.
-
October 13th, 2003, 08:07 PM
#10
avi123,
Since you already captured the int/double value in the form of string, why not scan for the decimal point or '.'. If there exists non-zero value character(s) after the decimal point, the string presentation must be a double.
-
October 13th, 2003, 09:44 PM
#11
Re: double and %
Originally posted by avi123
Hi,
I have a very big number (up to 15 digits) which is stored as a double
I need to know if this number is integer or not
but % function doesn't exist for double
I wanted to do chaeck myDoubleNum%1 is not equal to 0
any suggestions?
thanks
May I ask why it has to be 15 digits ?
Thanks a lot...
Regards,
-Nina
Last edited by hometown; October 13th, 2003 at 09:47 PM.
-
October 13th, 2003, 10:31 PM
#12
Re: Re: double and %
Originally posted by hometown
May I ask why it has to be 15 digits ?
Thanks a lot...
Regards,
-Nina
different program have differ specifications and requirements..In what he is doin he needs to manipulate a 15 didgit number...
or he simply wants to learn how to manipulate in case of a 15 digit no...
<edit>BTW : he said UPTO 15..</edit>
Last edited by Joseph_R_Thomas; October 13th, 2003 at 10:36 PM.
R. Thomas
"Be anxious for nothing, but in everything by prayer and supplication, with thanksgiving, let your requests be made know to God; and the peace of God, which surpasses all understanding, will guard your hearts and minds through Christ Jesus."Philippians 4:6-7
"Rejoice always, pray without ceasing, in everything give thanks; for this is the will of God in Christ Jesus for you."1Thess. 5:16-18
-
October 14th, 2003, 07:13 AM
#13
Instead of modf, use floor() instead. Do use the "epsilon" approach suggested by Yves. I like using "FLT_MIN" for that purpose.
Code:
#include <math.h>
#include <float.h>
int function(double dbl)
{
/* ... */
if ( ( dbl - floor(dbl) ) < FLT_MIN )
{
/* treat as integer. */
}
else
{
/* not integer. */
}
/* ... */
}
-rick
Last edited by rfmobile; October 14th, 2003 at 08:17 AM.
-
October 14th, 2003, 07:33 AM
#14
If my memory serves me right, this doesn't work for negative numbers.
Get this small utility to do basic syntax highlighting in vBulletin forums (like Codeguru) easily.
Supports C++ and VB out of the box, but can be configured for other languages.
-
October 14th, 2003, 08:17 AM
#15
Yup, I forgot the fabs().
-rick
Posting Permissions
- You may not post new threads
- You may not post replies
- You may not post attachments
- You may not edit your posts
-
Forum Rules
|
Click Here to Expand Forum to Full Width
|