Clojure版本:

(defn comb [k l]
  (if (= 1 k) (map vector l)
      (apply concat
             (map-indexed
              #(map (fn [x] (conj x %2))
                    (comb (dec k) (drop (inc %1) l)))
              l))))

现在又出现了祖辈COBOL，一种饱受诟病的语言。

让我们假设一个包含34个元素的数组，每个元素8个字节(完全是任意选择)。其思想是枚举所有可能的4元素组合，并将它们加载到一个数组中。

我们使用4个指标，每个指标代表4个组中的每个位置

数组是这样处理的:

    idx1 = 1
    idx2 = 2
    idx3 = 3
    idx4 = 4

我们把idx4从4变到最后。对于每个idx4，我们得到一个唯一的组合 四人一组。当idx4到达数组的末尾时，我们将idx3增加1，并将idx4设置为idx3+1。然后再次运行idx4到最后。我们以这种方式继续，分别增加idx3、idx2和idx1，直到idx1的位置距离数组末端小于4。算法就完成了。

1          --- pos.1
2          --- pos 2
3          --- pos 3
4          --- pos 4
5
6
7
etc.

第一次迭代:

1234
1235
1236
1237
1245
1246
1247
1256
1257
1267
etc.

一个COBOL的例子:

01  DATA_ARAY.
    05  FILLER     PIC X(8)    VALUE  "VALUE_01".
    05  FILLER     PIC X(8)    VALUE  "VALUE_02".
  etc.
01  ARAY_DATA    OCCURS 34.
    05  ARAY_ITEM       PIC X(8).

01  OUTPUT_ARAY   OCCURS  50000   PIC X(32).

01   MAX_NUM   PIC 99 COMP VALUE 34.

01  INDEXXES  COMP.
    05  IDX1            PIC 99.
    05  IDX2            PIC 99.
    05  IDX3            PIC 99.
    05  IDX4            PIC 99.
    05  OUT_IDX   PIC 9(9).

01  WHERE_TO_STOP_SEARCH          PIC 99  COMP.

* Stop the search when IDX1 is on the third last array element:

COMPUTE WHERE_TO_STOP_SEARCH = MAX_VALUE - 3     

MOVE 1 TO IDX1

PERFORM UNTIL IDX1 > WHERE_TO_STOP_SEARCH
   COMPUTE IDX2 = IDX1 + 1
   PERFORM UNTIL IDX2 > MAX_NUM
      COMPUTE IDX3 = IDX2 + 1
      PERFORM UNTIL IDX3 > MAX_NUM
         COMPUTE IDX4 = IDX3 + 1
         PERFORM UNTIL IDX4 > MAX_NUM
            ADD 1 TO OUT_IDX
            STRING  ARAY_ITEM(IDX1)
                    ARAY_ITEM(IDX2)
                    ARAY_ITEM(IDX3)
                    ARAY_ITEM(IDX4)
                    INTO OUTPUT_ARAY(OUT_IDX)
            ADD 1 TO IDX4
         END-PERFORM
         ADD 1 TO IDX3
      END-PERFORM
      ADD 1 TO IDX2
   END_PERFORM
   ADD 1 TO IDX1
END-PERFORM.

另一种python递归解决方案。

def combination_indicies(n, k, j = 0, stack = []):   
    if len(stack) == k:            
        yield list(stack)
        return
        
    for i in range(j, n):
        stack.append(i)
        for x in combination_indicies(n, k, i + 1, stack):            
            yield x
        stack.pop()  
        
list(combination_indicies(5, 3))

输出:

[[0, 1, 2],
 [0, 1, 3],
 [0, 1, 4],
 [0, 2, 3],
 [0, 2, 4],
 [0, 3, 4],
 [1, 2, 3],
 [1, 2, 4],
 [1, 3, 4],
 [2, 3, 4]]

《计算机编程艺术，卷4A:组合算法，第1部分》第7.2.1.3节中算法L(字典组合)的C代码:

#include <stdio.h>
#include <stdlib.h>

void visit(int* c, int t) 
{
  // for (int j = 1; j <= t; j++)
  for (int j = t; j > 0; j--)
    printf("%d ", c[j]);
  printf("\n");
}

int* initialize(int n, int t) 
{
  // c[0] not used
  int *c = (int*) malloc((t + 3) * sizeof(int));

  for (int j = 1; j <= t; j++)
    c[j] = j - 1;
  c[t+1] = n;
  c[t+2] = 0;
  return c;
}

void comb(int n, int t) 
{
  int *c = initialize(n, t);
  int j;

  for (;;) {
    visit(c, t);
    j = 1;
    while (c[j]+1 == c[j+1]) {
      c[j] = j - 1;
      ++j;
    }
    if (j > t) 
      return;
    ++c[j];
  }
  free(c);
}

int main(int argc, char *argv[])
{
  comb(5, 3);
  return 0;
}

作为迭代器对象实现的MetaTrader MQL4非常快速的组合。

代码很容易理解。

我对很多算法进行了基准测试，这个算法真的非常快——大约比大多数next_combination()函数快3倍。

class CombinationsIterator { private: int input_array[]; // 1 2 3 4 5 int index_array[]; // i j k int m_elements; // N int m_indices; // K public: CombinationsIterator(int &src_data[], int k) { m_indices = k; m_elements = ArraySize(src_data); ArrayCopy(input_array, src_data); ArrayResize(index_array, m_indices); // create initial combination (0..k-1) for (int i = 0; i < m_indices; i++) { index_array[i] = i; } } // https://stackoverflow.com/questions/5076695 // bool next_combination(int &item[], int k, int N) bool advance() { int N = m_elements; for (int i = m_indices - 1; i >= 0; --i) { if (index_array[i] < --N) { ++index_array[i]; for (int j = i + 1; j < m_indices; ++j) { index_array[j] = index_array[j - 1] + 1; } return true; } } return false; } void getItems(int &items[]) { // fill items[] from input array for (int i = 0; i < m_indices; i++) { items[i] = input_array[index_array[i]]; } } };

测试上述迭代器类的驱动程序:

//+------------------------------------------------------------------+ //| | //+------------------------------------------------------------------+ // driver program to test above class #define N 5 #define K 3 void OnStart() { int myset[N] = {1, 2, 3, 4, 5}; int items[K]; CombinationsIterator comboIt(myset, K); do { comboIt.getItems(items); printf("%s", ArrayToString(items)); } while (comboIt.advance()); }

输出: 1 2 3 1 2 4 1 2 5 1 3 4 1 3 5 1 4 5 2 3 4 2 3 5 2 4 5 3 4 5

不需要进行集合操作。这个问题几乎和循环K个嵌套循环一样，但你必须小心索引和边界(忽略Java和OOP的东西):

 public class CombinationsGen {
    private final int n;
    private final int k;
    private int[] buf;

    public CombinationsGen(int n, int k) {
        this.n = n;
        this.k = k;
    }

    public void combine(Consumer<int[]> consumer) {
        buf = new int[k];
        rec(0, 0, consumer);
    }

    private void rec(int index, int next, Consumer<int[]> consumer) {
        int max = n - index;

        if (index == k - 1) {
            for (int i = 0; i < max && next < n; i++) {
                buf[index] = next;
                next++;
                consumer.accept(buf);
            }
        } else {
            for (int i = 0; i < max && next + index < n; i++) {
                buf[index] = next;
                next++;
                rec(index + 1, next, consumer);
            }
        }
    }
}

像这样使用:

 CombinationsGen gen = new CombinationsGen(5, 2);

 AtomicInteger total = new AtomicInteger();
 gen.combine(arr -> {
     System.out.println(Arrays.toString(arr));
     total.incrementAndGet();
 });
 System.out.println(total);

获得预期的结果:

[0, 1]
[0, 2]
[0, 3]
[0, 4]
[1, 2]
[1, 3]
[1, 4]
[2, 3]
[2, 4]
[3, 4]
10

最后，将索引映射到您可能拥有的任何数据集。