Shell Programming and Scripting

I am looking for a better way to match real numbers within a specified tolerance range. My current code is as follows:

Code:

      if ($1 !~ /^CASE/) for(i=1;i in G;i++) if (G[i] >= $5-1  && G[i] <= $5+1)
         { print $1,$4,$5,J[i],G[i] }
           else { print $1,"NO MATCH" }

where $5 and G[i] are real numbers e.g. "12.456" and "11.543". Right now the code prints out several numbers within the range of 1+ 12.456 and 1-12.456.

Is there a better way of accomplishing this? I would like to get the closest number without manually having to adjust the tolerance range of 1.

Thanks

Hi.

See Relative difference - Wikipedia, the free encyclopedia ... cheers, drl

What drl is telling you:
1. floating point numbers represent values correctly only to a certain level of precision
so the = part of a comparison does not work like you would think.

2. awk uses double precision floating point number for aritmetic operations, so #1 applies.

3. in /usr/include/math.h you will find macros defined:

Code:

int isgreater(real-floating x, real-floating y);
int isgreaterequal(real-floating x, real-floating y);
int isless(real-floating x, real-floating y);
int islessequal(real-floating x, real-floating y);
int islessgreater(real-floating x, real-floating y);

these are the standard way of doing floating point comparisons on your system. Notice there is no "equal()", for reasons just explained.

Here is a good explanation of FP numbers -- the nearly universal IEEE-754 format:
IEEE Standard 754 Floating-Point

How would I apply this to my code? I can't seem to view any of the links right now.