Previous fileTop of DocumentContentsIndex pageNext file
Apache C++ Standard Library User's Guide

14.2 Sorting Algorithms


NOTE -- The example programs described in the following sections have been combined and are included in the file alg7.cpp. As in Chapter13, we generally omit output statements from the descriptions of the programs provided here, although they are included in the complete versions.

The C++ Standard Library provides two fundamental sorting algorithms, described as follows:

The std::sort() algorithm is slightly faster, but it does not guarantee that equal elements in the original sequence retain their relative orderings in the final result. If order is important, the std::stable_sort() version should be used.

Because these algorithms require random access iterators, they can be used only with vectors, deques, and ordinary C pointers. Note, however, that the list container provides its own sort() member function.

The comparison operator can be explicitly provided when the default operator < is not appropriate. This is used in the example program to sort a list into descending, rather than ascending order. An alternative technique for sorting an entire collection in the inverse direction is to describe the sequence using reverse iterators.


NOTE -- Another sorting algorithm is provided by the heap operations described in Section 14.7.

The following example program illustrates the std::sort() algorithm being applied to a vector, and the std::sort() algorithm with an explicit comparison operator being used with a deque.

14.2.1 Partial Sort

The generic algorithm std::partial_sort() sorts only a portion of a sequence. In the first version of the algorithm, three iterators are used to describe the beginning, middle, and end of a sequence. If n represents the number of elements between the start and middle, then the smallest n elements are moved into this range in order. The remaining elements are moved into the second region. The order of the elements in this second region is undefined.

A second version of the algorithm leaves the input unchanged. The output area is described by a pair of random access iterators. If n represents the size of this area, the smallest n elements in the input are moved into the output in order. If n is larger than the input, the entire input is sorted and placed in the first n locations in the output. In either case, the end of the output sequence is returned as the result of the operation.

Because the input to this version of the algorithm is specified only as a pair of input iterators, the std::partial_sort_copy() algorithm can be used with any of the containers in the C++ Standard Library. In the example program, it is used with a list:



Previous fileTop of DocumentContentsIndex pageNext file