Solving Error in C++ Digit Algorithm Program

[chaoseverlasting]

I wrote a program to find the number of digits of an integer, but I always get the wrong result. Could someone point out the error? Here's the source code:

#include <iostream>
#include <cmath>

using namespace std;

int main()
{
int i=53443;
count<<log(i);
return 0;
}


The log function gives me a weird result each time. I don't know why though. I am using dev c++ compiler...

 

[D H]

The log function computes the natural logarithm.

 

[Borek]

man log10

 

[D H]

Is this homework? I don't want to go much further with help if it is.

 

[chaoseverlasting]

This isn't homework. I am trying to write a program which needs this bit. I can't seem to figure this out. So is it log(e) or log (10) ?

 

[chaoseverlasting]

If that's it (the natural log- base 10 difference), I can figure the rest out. Is there perhaps some property of integers and the way this function works (the taylor series of log(1+x) maybe?) that's causing this?

 

[D H]

RTFM, or as Borek said, man log10 (and man log won't hurt). If you are not on a UNIX machine, there is always the internet. What does log() compute?

BTW, neither log() nor log10() is really what you want. Will this approach work on negative integers or on 64 bit (long long) integers?

 

[zeitghost]

Personally I'd use sprintf & count the characters in the buffer...

 

[D H]

Since this is not homework, I will give my solution. Keep dividing by ten (the integer, not the double) until the result becomes zero. This works for all sizes of integers, negative integers, is much faster and more accurate than using log10(), and is much, much faster than sprintf.

int ndigits = 0;
for (same_int_type_as_i rem = i; rem != 0; rem /= 10) {
   ++ndigits;
}

 

[Borek]

When I went to bed yesterday I thought of strlen(itoa(abs(i))) but I am not addicted to the web enough to post such things immediately This is similar to sprintf, just much easier to code.

 

[chaoseverlasting]

When I went to bed yesterday I thought of strlen(itoa(abs(i))) but I am not addicted to the web enough to post such things immediately This is similar to sprintf, just much easier to code.

Since this is not homework, I will give my solution. Keep dividing by ten (the integer, not the double) until the result becomes zero. This works for all sizes of integers, negative integers, is much faster and more accurate than using log10(), and is much, much faster than sprintf.

int ndigits = 0;
for (same_int_type_as_i rem = i; rem != 0; rem /= 10) {
   ++ndigits;
}

Thank you. That helps .