problem with floating point number loops

Shell Programming and Scripting
1

Hey,

I guess I am just to stupid and am not seeing the "wood for the trees", but I am always getting strange errors.

I want to create a mesh with coordinates like:

x y z
3.1 3.0 0.75 0 0 1
3.1 2.9 0.75 0 0 1
3.1 2.8 0.75 0 0 1
3.1 2.7 0.75 0 0 1
3.0 3.0 0.75 0 0 1
3.0 2.9 0.75 0 0 1
3.0 2.8 0.75 0 0 1
3.0 2.7 0.75 0 0 1

Here is my skript:

echo "Abstand Berechnungspunkte in Meter="
read abstand

z=0.75
a=0
b=0
c=1

for (( i=$abstand; i<0; i=$i+$abstand )) 
do
			for (( h=$abstand; h<4.61; h=$h+$abstand )) 
			do
						x=`echo "-3.616+$i" |bc -l `
						y=`echo "-0.085+$h" |bc -l `
						print $x " " $y " " $z " " $a " " $b " " $c
			done
done

or I also tried this one:

echo "Abstand Berechnungspunkte in Meter="
read abstand 

    z=.75
    a=0
    b=0
    c=1
    ab=$abstand
    i=-3.616
    h=0
    maxx=0
    maxy=4.61
    
    while [ $i -le $maxx ]
    do 
            x=`echo "-3.616+$i" |bc -l`
            i=`echo "$i+$ab" |bc -l`
    done

But it is always creating the same problem regarding numbers with commata. Becasue my input is e.g., 0.02
line 11: ((: i=0.1: syntax error: invalid arithmetic operator (error token is ".1")

Does anyone has an idea? Thanks for any help and sorry for the stupid question,
Sam

2

The shell will not allow you to loop using a floating point value. As a general rule floating point values should not be used in loop counters even when the language allows it.

3

ok. thanks.

Do you then have an idea how to realize such a grid by a loop easily?

And thanks for your reply, sam

4

A possible work-around: convert all meter into centimeter values, e.g. by means of sed or awk :o

5

My Deutsch isn't all that great, so to make sure I'm clear on what you're doing here - you want to find the number of steps between between two constant values.

Isn't this essentially a division problem? For example if the abstand=2, your xmax is 6 and your ymax is 4 the output table would be:

6 4 0.75 0 0 1
6 2 0.75 0 0 1
6 0 0.75 0 0 1
4 4 0.75 0 0 1
4 2 0.75 0 0 1
4 0 0.75 0 0 1
2 4 0.75 0 0 1
2 2 0.75 0 0 1
2 0 0.75 0 0 1
0 4 0.75 0 0 1
0 2 0.75 0 0 1
0 0 0.75 0 0 1

Also, are you limited to using a shell script? The only way I know of to do floating point comparisons in bash is pretty kludgy. If you have perl available and my understanding is correct I could definitely create a script to handle this.

This would only push the failure point to millimeters.

6

Even floating point has a failure point. The question is how much accuracy is necessary.

7

Math::BigFloat could express atoms in meters, so I figured that wouldn't be an issue. Likewise, I assumed that more than a couple of decimal places were necessary. If this is just for a limited degree of accuracy an expr or bc kludge would be sufficient, but I wouldn't want to maintain it.

8

thanks, right. I am stupid...

But there is another problem. Do you know how the right syntax has to eb for such a for-loop in bash. Because the while-loop doesn't work equally.

And I am always getting errors. Do you have any idea how to transfer maybe my for-loop into another loop? Or do I have to use "["?

Thanks for help,
Sam

9

Here is how to do it in ksh93.

#!/bin/ksh93

float abstand=0.1
float i
float h

for (( i=3.1; i >= 3.0; i-=abstand ))
do
    for (( h=3.0; h >= 2.7; h-=abstand ))
    do
        printf "%1.1f %1.1f 0.75 0 0 1\n" i h
    done
done

Output is

3.1 3.0 0.75 0 0 1
3.1 2.9 0.75 0 0 1
3.1 2.8 0.75 0 0 1
3.1 2.7 0.75 0 0 1
3.0 3.0 0.75 0 0 1
3.0 2.9 0.75 0 0 1
3.0 2.8 0.75 0 0 1
3.0 2.7 0.75 0 0 1
10

ok. but the problem is, i am only having mksh and the I am never able to get the AT&T ksh93, because of their license problems....

So Do you know how the syntax should be for a for-loop in mksh? (mksh even doesn't know float then "for ((")

Thanks again for your help

11

What license problems?