[Source]
void bubblesort(char skaters[][15], float averages[], int count)
{
float stopt={0};
char *stop={0},*names[10]={0};
int a={0},b={0};
for(count=0;count<10;count++)
{
names[count] = &skaters[count][0];
}
for(a=0;a<10-1;a++)
{
for(b=a+1;b<10;b++)
{
if(averages[a]<averages);
{
stop=names[a];
stopt=averages[a];
names[a]=names;
averages[a]=averages;
names=stop;
averages=stopt;
}
}
}
}
[/Source]How to bubble sort floats?
Author
Hello Everyone!
I can't bubble sort floating point numbers, I think I need to use pointers:
can you atleast explain what is what here ? I can hardly understand what these variables are.
Syntax Error:
Your extra ";" effectively makes your if() a pointless construct (my compiler warns me of this, maybe you should double check you've enabled warnings and not ignored this one!). Because of this, the swapping done between { and } will be executed for every iteration - generally not a lot of use to you [smile].
Something you may not be aware of is that comparing two floating point numbers is a bit of a grey area.
Normal bubble sort relies on lines like if(averages[a]<averages) which, for small floats (or very similar floats) may not give the results you expect. For better results you want to compare them to an epsilon value.
I think there are some definitions of what epsilon should be, but its been a while since I've messed about with it (the number varies for comparing large floats or comparing very small floats iirc).
For example, equality should be something like if( fabs(averages[a] - averages) < 0.000001f ). I think I saw a C-Macro that worked for LEQUAL/GEQUAL based on the signed bit in the representation, but as I said - it's been a while [smile]
hth
Jack
if(averages[a]<averages); // <<--- the ';' should not be here! { stop=names[a]; stopt=averages[a]; names[a]=names; averages[a]=averages; names=stop; averages=stopt; }Your extra ";" effectively makes your if() a pointless construct (my compiler warns me of this, maybe you should double check you've enabled warnings and not ignored this one!). Because of this, the swapping done between { and } will be executed for every iteration - generally not a lot of use to you [smile].
Quote:I can't bubble sort floating point numbers
Something you may not be aware of is that comparing two floating point numbers is a bit of a grey area.
Normal bubble sort relies on lines like if(averages[a]<averages) which, for small floats (or very similar floats) may not give the results you expect. For better results you want to compare them to an epsilon value.
I think there are some definitions of what epsilon should be, but its been a while since I've messed about with it (the number varies for comparing large floats or comparing very small floats iirc).
For example, equality should be something like if( fabs(averages[a] - averages) < 0.000001f ). I think I saw a C-Macro that worked for LEQUAL/GEQUAL based on the signed bit in the representation, but as I said - it's been a while [smile]
hth
Jack
Don't feel bad about using the STL:
#include <iostream>#include <algorithm>int main() { float a[7] = {23.4F, 1.5F, 79.31F, -1.4F, 1.2F, 1.73F, 1.7F}; std::sort(a, a+7); // sort 7 elements for (int i=0; i<7; i++) { std::cout << a << "\n"; } std::cin.get(); return 0;}This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement
