Sorting an array of unique random numbers at insertion
CodeGuru Home VC++ / MFC / C++ .NET / C# Visual Basic VB Forums Developer.com
Results 1 to 8 of 8

Thread: Sorting an array of unique random numbers at insertion

  1. #1
    Join Date
    Sep 2011
    Posts
    1

    Exclamation Sorting an array of unique random numbers at insertion

    I have a piece of code that works well to fill an array with a unique random numbers. My problem now is, that I want to sort these numbers, but not after the array is full but as new numbers are being inserted. So as each new number is inserted into the array, if finds the position it is meant to be in. The code I have for creating unique random numbers is below, thank you in advance:

    #include <iostream>
    #include <stdio.h>
    #include <stdlib.h>
    #include <time.h>
    using namespace std;

    #define MAX 2000 // Values will be in the range (1 .. MAX)
    static int seen[MAX]; // These are automatically initialised to zero
    // by the compiler because they are static.
    static int randomNum[1000];

    int main (void) {
    int i;

    srand(time(NULL)); // Seed the random number generator.

    for (i=0; i<1000; i++)
    {
    int r;
    do
    {
    r = rand() / (RAND_MAX / MAX + 1);
    }
    while (seen[r]);
    seen[r] = 1;
    randomNum[i] = r + 1;
    }

    for (i=0; i<1000; i++)
    cout << randomNum[i] << endl;

    return 0;
    }

  2. #2
    Lindley is offline Elite Member Power Poster
    Join Date
    Oct 2007
    Location
    Fairfax, VA
    Posts
    10,891

    Re: Sorting an array of unique random numbers at insertion

    If you're looking for efficiency, a std::set may be more suitable than an array.

  3. #3
    Join Date
    Aug 2000
    Location
    New York, NY, USA
    Posts
    5,527

    Re: Sorting an array of unique random numbers at insertion

    Quote Originally Posted by woodman_102 View Post
    ...I want to sort these numbers, but not after the array is full but as new numbers are being inserted.
    What is the difference?
    Anyway, instead of assigning your new random number to the randomNum[i], you could loop from randomNum[i] to randomNum[0] looking for the insertion point. If the new number is greater assign; if it is less move last element up and keep going.
    Vlad - MS MVP [2007 - 2012] - www.FeinSoftware.com
    Convenience and productivity tools for Microsoft Visual Studio:
    FeinViewer - an integrated GDI objects viewer for Visual C++ Debugger, and more...

  4. #4
    Join Date
    May 2009
    Posts
    2,413

    Re: Sorting an array of unique random numbers at insertion

    Quote Originally Posted by Lindley View Post
    If you're looking for efficiency, a std::set may be more suitable than an array.
    Not just for efficiency reasons. Std::set has the two properties the OP is asking for - elements are unique and they're kept sorted at all times.

    So it's just to keep inserting random ints until the set reaches the wanted size. Whenever the set is iterated the ints will appear in sorted order (and they will be unique).
    Last edited by nuzzle; September 30th, 2011 at 03:17 PM.

  5. #5
    Lindley is offline Elite Member Power Poster
    Join Date
    Oct 2007
    Location
    Fairfax, VA
    Posts
    10,891

    Re: Sorting an array of unique random numbers at insertion

    Quote Originally Posted by VladimirF View Post
    What is the difference?
    Well, running time of course. Assuming a naive insertion sort is used, keeping the data sorted after every insertion results in a running time of O(n*(n+log n)) at best. Sorting once after all insertions are done is O(n + n*log n), much faster.

    If a std::set is used, keeping the data sorted after every insertion is O(n*log n) but in some cases cache locality may not be as good.

    If an array is required, "library sort" (basically insertion sort where some empty slots are left between valid elements) can be done in O(n*log n) but in practice it's still slower.
    Last edited by Lindley; September 30th, 2011 at 03:36 PM.

  6. #6
    Join Date
    May 2009
    Posts
    2,413

    Re: Sorting an array of unique random numbers at insertion

    Quote Originally Posted by woodman_102 View Post
    So as each new number is inserted into the array, if finds the position it is meant to be in. The code I have for creating unique random numbers is below, thank you in advance:
    Note that strictly the randomNum[] array isn't necessary since seen[] already holds all information about which ints have been randomly selected. Also seen[] is inherently sorted. Each newly selected int will be at the "position it is meant to be in" because the position is the int.
    Last edited by nuzzle; September 30th, 2011 at 04:35 PM.

  7. #7
    Join Date
    May 2009
    Posts
    2,413

    Re: Sorting an array of unique random numbers at insertion

    Quote Originally Posted by Lindley View Post
    Well, running time of course. Assuming a naive insertion sort is used, keeping the data sorted after every insertion results in a running time of O(n*(n+log n)) at best. Sorting once after all insertions are done is O(n + n*log n), much faster.

    If a std::set is used, keeping the data sorted after every insertion is O(n*log n) but in some cases cache locality may not be as good.

    If an array is required, "library sort" (basically insertion sort where some empty slots are left between valid elements) can be done in O(n*log n) but in practice it's still slower.
    That's not correct.

    O(n*(n+log n)) and O(n + n*log n) are not even proper complexity measures. It's O(n*n) and O(n*log n) respectively.

    In the first case, even if you use an O(log n) binary search to locate the insertion point you need to insert it which is an O(n) operation. Repeating n insertions gives a complexity of O(n * n).

    In the second case you place the items in the array in arbitrary order which is an O(n) operation. Afterwards you sort at O(n* log n). Taken together this gives O(n * log n) complexity.

    Inserting all items in an ordered set has O(n * log n) complexity as you stated.

  8. #8
    Join Date
    Jun 2009
    Location
    France
    Posts
    2,318

    Re: Sorting an array of unique random numbers at insertion

    Some would argue that when it comes to reads, a sorted vector is faster than a set, because of locality (I believe Meyers mentions it too in one of his "more effective"). This is an adaptor that transforms the vector's interface into that of a set.

    That said, it has methods specifically dedicated to NOT sort after every insertion, given its higher price.
    Is your question related to IO?
    Read this C++ FAQ LITE article at parashift by Marshall Cline. In particular points 1-6.
    It will explain how to correctly deal with IO, how to validate input, and why you shouldn't count on "while(!in.eof())". And it always makes for excellent reading.

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  


Windows Mobile Development Center


Click Here to Expand Forum to Full Width

This is a CodeGuru survey question.


Featured


HTML5 Development Center