2

I've got a quick and I am assuming question but I have not been able to find anything online.

How to calculates the average of elements in an unsigned char array? Or more like it, perform operations on an unsigned char?

Nawaz
  • 353,942
  • 115
  • 666
  • 851
Everton
  • 109
  • 2
  • 9
  • 4
    Do you know how to take the average of a data set? Do you understand the concept of arrays? Do you want to understand the syntax for working with arrays? Do you want to do it in C or C++? (**C and C++ are different languages.** You will get different answers for each. The `[c]` and `[c++]` tags are not equivalent!) – In silico Feb 10 '11 at 04:10

3 Answers3

5

C++03 and C++0x:

#include <numeric>

int count = sizeof(arr)/sizeof(arr[0]);
int sum = std::accumulate<unsigned char*, int>(arr,arr + count,0);
double average = (double)sum/count;

Online Demo : http://www.ideone.com/2YXaT

C++0x Only (using lambda)

#include <algorithm>

int sum = 0;
std::for_each(arr,arr+count,[&](int n){ sum += n; });
double average = (double)sum/count;

Online Demo : http://www.ideone.com/IGfht

Nawaz
  • 353,942
  • 115
  • 666
  • 851
1

About like with anything else, you add them up and divide by the count. To avoid overflow, you'll typically want to convert them to something larger while you're doing the math. If (as is common) you want a floating point result, you'll want to do all the math on floating point as well.

Jerry Coffin
  • 476,176
  • 80
  • 629
  • 1,111
  • 2
    No need to do all the math as floating point. You can wait to convert until you finish the sum. If your integers are 32-bit, you'd need a 24MB array to have any chance of overflow, so I think it's perfectly reasonable to keep the sum in an `unsigned int`. – R.. GitHub STOP HELPING ICE Feb 10 '11 at 04:20
  • (And if there is a chance of overflow, you're going to have a hard time getting the code right with floating point either due to loss of precision!) – R.. GitHub STOP HELPING ICE Feb 10 '11 at 04:21
  • @R.: Yes, for the job exactly as stated, a 32-bit int has a pretty decent chance of being adequate -- but, 1) it won't always be, and 2) doing so is *relatively* fragile, so (for example) a seemingly minor change in the input type can break "known working" code. At one time using floating point was expensive enough that it was worth avoiding whenever possible, but that's changed. In fact, some CPUs now do some integer operations by converting the data to FP, doing the operation, the converting the result back to an integer. Clearly in such a case, using FP isn't a major problem... – Jerry Coffin Feb 10 '11 at 08:34
  • If nothing else, you need to make it clear that you mean `double`, not `float`, since `float` will have worse overflow issues (after a large number of elements, around 32k, are added, adding further elements **will not change the sum**!) which are harder to detect and harder to work around. IMO, using floating point any time you don't really want its semantics is a huge programming error that's sure to bite you unless you're an expert in numerical analysis. – R.. GitHub STOP HELPING ICE Feb 10 '11 at 10:36
1

Arithmetic operations work just fine on unsigned char, although you may occasionally be surprised by the fact that arithmetic in C always promotes to int.

In C++'s Standard Template Library,

#include <numeric>
template<class InputIterator, class T>
T accumulate(InputIterator first, InputIterator last, T init);

To calculate the sum of unsigned char arr[], you may use accumulate(arr, arr + sizeof(arr) / sizeof(arr[0]), 0). (0 is an int here. You may find it more appropriate to use a different type.)

Without STL, this is trivially computed with a loop.

The average is the sum divided by the length (sizeof(arr) / sizeof(arr[0])).

ephemient
  • 198,619
  • 38
  • 280
  • 391
  • How can I hold an unsigned char value? Let's say sum += unSignedCharArray[i][j]. what is the type of sum? – Everton Feb 10 '11 at 05:10
  • @Everton: See my complete solution. the sum is stored in an `int` type variable! – Nawaz Feb 10 '11 at 05:21
  • Can you show me how to implement without STL? I am a beginner and it would help me a lot! Thank you! – Everton Feb 10 '11 at 07:49