Thread: Question about a certain code with the C++ syntax

Wow it fixed the majority of the problem! Now I dropped 4 errors down to just one it looks like. Visual Basic agrees with Dev C++ Bloodshead.

So apparently sometimes if you store to many variables in one function it can mess up which I suspect what promoted that error. If I am correct?

Code:

for (int j=0; j<numberofterms-number_of_elements_avg; j++)        {
            // j - number_of_elements_avg/2.0 
            unsigned int start = j - number_of_elements_avg/2;
            average = accumulate(temperature.begin() + start, temperature.begin() + j + number_of_elements_avg/2 + 1, 0.0) / (number_of_elements_avg + 1);
        }

Now...the one error ONLY.

1>------ Build started: Project: RunningMean, Configuration: Debug Win32 ------
1> RunningMean.cpp
1>c:\users\futurenws4caster\documents\visual studio 2010\projects\runningmean\runningmean\RunningMean. h(50): warning C4244: '=' : conversion from 'double' to 'float', possible loss of data
1> RunningMean.vcxproj -> c:\users\futurenws4caster\documents\visual studio 2010\Projects\RunningMean\Debug\RunningMean.dll
========== Build: 1 succeeded, 0 failed, 0 up-to-date, 0 skipped ==========

I would recommend that your change your "float" to "double", you will get several benefits, more precession an less round off problems. Today in my opinion the use of the float data type for desktop applications should be avoided. There are situations where use of float is useful but in most cases using doubles is much better.

The other option would be to make sure the last parameter is a float by using 0.0f or casting it to a float. Also I noticed that you took part of the calculations out of the function call I recommend you think about finishing it up by replacing the other calculation as well.

Jim

Originally Posted by jimblumberg

Today in my opinion the use of the float data type for desktop applications should be avoided. There are situations where use of float is useful but in most cases using doubles is much better.

That's new to me, thanks.
Where (&& why) do you think, "use of float is useful" ?
(I'm learning OpenGL now, and every text seems to use float, though.)

By the way in addition on Dev Blooshead C++ why doesnt the #include "stdafx.h" work?

Interesting where can I find more info about that? Any links Jim?

O_o

That statement was too broad to resolve to a useful "rule of thumb".

Use a type that is appropriate to the precision necessary versus the performance of the hardware with variables of that precision.

For graphics using "OpenGL" or "Direct3D" you should usually be using single-precision even where the API offers double-precision possibilities because most hardware doesn't do anything with the extra precision while still costing a little extra time do to the conversion. Where the hardware does honor the request for double-precision calculations it is usually much slower than single-precision which is sufficient for the majority of rendering. Finally, even hardware that does support double-precision for some parts of the API don't support it the way you may think which can cause rendering problems.

Soma

Now that I got the code running. Here's what the executable is displaying. The bold is the input.

Enter the number of terms you want:
4
Enter how many point average you want take:
3
Enter Temp input in C 1: 20

AND THEN IT CRASHES AT THAT POINT. Why isnt it going through the for loop?

Code:

//Running Mean for 3 points with even weights
 
#include <iostream>
#include <iomanip>
#include <vector>
#include <numeric>
#include <functional>
using namespace std;
 
int main ()
{
    int numberofterms;
    int number_of_elements_avg; 
    float tempstorage;
    vector<float> temperature; //Define a vector of 5 floats
    vector<float> movingavg;
    int i; //Loop counter
 
    cout<<"Enter the number of terms you want: " <<endl;
    cin>>numberofterms;
    cout<<"Enter how many pt average you would like to take: "<<endl;
    cin>>number_of_elements_avg;
 
 
    for (i=0; i<numberofterms; i++)
    {  
        float average;
 
        cout<<"Enter temperature you want to input in Celsius "<<(i+1);
        cout<<": ";
        cin>>tempstorage;
        temperature.push_back(tempstorage);
 
        if (i=0)
        {
            movingavg[i]=temperature[i];
            cout<<"Temperature [1]= "<<temperature[0]<<" C"<<endl;
            continue;
        }
        else if (i=numberofterms-1)
        {
            movingavg[i]=temperature[i];
            cout<<"Temperature ["<<numberofterms<<"]= "<<temperature[numberofterms-1]<<" C"<<endl;
            continue;
        }
        for (int j=0; j<numberofterms-number_of_elements_avg; j++)
        {
            // j - number_of_elements_avg/2.0 
            unsigned int start = j - number_of_elements_avg/2;
            average = accumulate(temperature.begin() + start, temperature.begin() + j + number_of_elements_avg/2 + 1, 0.0) / (number_of_elements_avg + 1);
        }
        //NOT PART OF THE CODE average= (temperature[index-1]+temperature[index]+temperature[index+1]/3);
        //NOT PART OF THE CODE movingavg[index]=average;
        cout<<"Temperature ["<<i<<"] = "<<average<<" C"<<endl;
        cout<<"The original temperature ["<< i+1<<"]"<<tempstorage<<"] C has been replaced." <<endl;
    }
 
    system ("PAUSE");
    return 0;
}

Because you vector movingavg has no elements. You can't use assignment with an empty vector.

Jim

Originally Posted by jimblumberg

Because you vector movingavg has no elements. You can't use assignment with an empty vector.

Jim

I thought the movinavg has elements because its dependent on the temperature on this loop part:

Code:

 if (i=0)
        {            movingavg[i]=temperature[i];            cout<<"Temperature [1]= "<<temperature[0]<<" C"<<endl;            continue;        }        else if (i=numberofterms-1)        {            movingavg[i]=temperature[i];            cout<<"Temperature ["<<numberofterms<<"]= "<<temperature[numberofterms-1]<<" C"<<endl; 
            continue;

I thought the movinavg has elements because its dependent on the temperature on this loop part:

That is not how vector works, as you have been told. Unless you use pointers (and I am not telling you to use them) there is no such thing as a vector dependent on another vector. If you construct two vectors you must insert elements into both properly. The [] operator is merely for accessing elements. In my opinion the two most helpful modifiers for vector are vector::push_back() for adding one element to the end, and vector::insert() for adding many elements at once starting from a certain place. You can click on the red to see the appropriate references.


hth

I thought the movinavg has elements because its dependent on the temperature on this loop part:

As I have already said, several times... Re-read posts 5, 7, 11 and 53.....

You can't use assignment with an empty vector.

Jim

Originally Posted by Rod Micael

I thought the movinavg has elements because its dependent on the temperature on this loop part:

Code:

if (i=0)
        {            movingavg[i]=temperature[i];

Think post #11 said it best
Question about a certain code with the C++ syntax

To repeat something else Jim said - it's far far easier to figure out what's wrong if you use a debugger. I see you've built your code with both Visual Studio and GCC -- both Visual Studio and gdb would have been able to tell you the exact line with the problem.
If you get friendly with a debugger, you'll spend a lot less time being confounded by weird logic errors. Honestly, it's worth the time.

Couple of other comments:

Code:

        if (i=0)
....
        else if (i=numberofterms-1)

These should be "==" not "=".

I'm a bit confused about what you want your program to do. Say the user wants to input 5 values. Do you intend them all to be input, then the moving averages are computed after input is finished? Or do you want to calculate the moving averages on the fly as the user enters the data?

This loop looks like it's meant to calculate averages for all the data:

Code:

for (int j=0; j<numberofterms-number_of_elements_avg; j++)
{
     // j - number_of_elements_avg/2.0 
     unsigned int start = j - number_of_elements_avg/2;
     average = accumulate(temperature.begin() + start, temperature.begin() + j + number_of_elements_avg/2 + 1, 0.0) / (number_of_elements_avg + 1);
}

but this loop is nested inside your for loop for input, so will be run every time the user enters a number. If you want to do it 'on the fly' you don't need a second loop: as soon as you have number_of_elements_avg elements you can do an accumulate using 'i' to work out the start and end. However that'd mean you'd get (if I put in 10,20,30,40,50 for 3 avg points):

Code:

element  value  average
1             10       10
2             20       20
3             30       (10+20+30)/3
4             40       (20+30+40)/3
5             50       (30+40+50)/3

and it looks like you want

Code:

element  value  average
1             10       10
2             20       (10+20+30)/3
3             30       (20+30+40)/3
4             40       (30+40+50)/3
5             50       50

In which case you do need a second loop, but you don't want it inside your input loop.

Once you've fixed your array allocation I think you'll get another crash:

Code:

for (int j=0; j<numberofterms-number_of_elements_avg; j++)
        {
            // j - number_of_elements_avg/2.0 
            unsigned int start = j - number_of_elements_avg/2;

If j=0 and number_of_elements_avg=3, start is 0 - (3/2) = 0-1 = -1. Since start is unsigned, you'll get the unsigned representation of the signed number -1, which is probably around about 4 billion. Not going to work!

You could change the initial value of j, or change the calculation of start and end.

I also recommend that you also write a small program to experiment with vectors and try using your debugger to debug the problems you find, and if you don't find any problems, create some. The sooner you figure out your debugger the better. Without knowing how to use your debugger you will never become even a semi-accomplished programmer.

You also need to learn to read and understand the documentation for the features you are trying to use. Today, there is a wealth of information available to aid in the understanding of the standard features. If one web site doesn't explain the feature to your satisfaction, try another, or another until you exhaust the possibilities. After you exhaust the possibilities and you come to ask questions be sure you read and understand the answers provided. If you don't understand the answer, say so. Maybe we can restate the answer in another way that you will understand. But as it is you seem to either not be understanding the answers, or not learning from your mistakes.

Jim