【Leetcode】Course Schedule II #210

There are a total of n courses you have to take, labeled from 0 to n - 1.
Some courses may have prerequisites, for example to take course 0 you have to first take course 1, which is expressed as a pair: [0,1]
Given the total number of courses and a list of prerequisite pairs, return the ordering of courses you should take to finish all courses.
There may be multiple correct orders, you just need to return one of them. If it is impossible to finish all courses, return an empty array.
For example:
2, [[1,0]]
There are a total of 2 courses to take. To take course 1 you should have finished course 0. So the correct course order is [0,1]
4, [[1,0],[2,0],[3,1],[3,2]]

There are a total of 4 courses to take. To take course 3 you should have finished both courses 1 and 2. Both courses 1 and 2 should be taken after you finished course 0. So one correct course order is [0,1,2,3]. Another correct ordering is[0,2,1,3].

url: https://leetcode.com/problems/course-schedule-ii/

一道经典的拓扑排序题，代码如下：

<p class="p1"><span class="s1">class</span><span class="s2"> Solution {</span></p><p class="p2"><span class="s2">public</span><span class="s3">:</span></p><p class="p1"><span class="s2">  </span><span class="s1">enum</span><span class="s2"> Status</span></p><p class="p1"><span class="s2">  {</span></p><p class="p1"><span class="s2">    unvisited = </span><span class="s4">0</span><span class="s2">,</span></p><p class="p1"><span class="s2">    visited,</span></p><p class="p1"><span class="s2">    temp_visited</span></p><p class="p1"><span class="s2">  };</span></p><p class="p3"><span class="s2">  </span></p><p class="p1"><span class="s2">  </span><span class="s5">vector</span><span class="s2"><</span><span class="s1">int</span><span class="s2">> findOrder(</span><span class="s1">int</span><span class="s2"> numCourses, </span><span class="s5">vector</span><span class="s2"><</span><span class="s5">pair</span><span class="s2"><</span><span class="s1">int</span><span class="s2">, </span><span class="s1">int</span><span class="s2">>>& prerequisites) {</span></p><p class="p1"><span class="s2">    </span><span class="s6">Status</span><span class="s2"> *record = </span><span class="s1">new</span><span class="s2"> </span><span class="s6">Status</span><span class="s2">[numCourses];</span></p><p class="p1"><span class="s2">    </span><span class="s1">for</span><span class="s2"> (</span><span class="s1">int</span><span class="s2"> i = </span><span class="s4">0</span><span class="s2">; i < numCourses; i++) {</span></p><p class="p1"><span class="s2">      record[i] = </span><span class="s6">unvisited</span><span class="s2">;</span></p><p class="p1"><span class="s2">    }</span></p><p class="p1"><span class="s2">    </span><span class="s5">vector</span><span class="s2"><</span><span class="s1">int</span><span class="s2">> result;</span></p><p class="p3"><span class="s2">    </span></p><p class="p4"><span class="s3">    </span><span class="s2">unordered_map</span><span class="s3"><</span><span class="s1">int</span><span class="s3">, </span><span class="s2">vector</span><span class="s3"><</span><span class="s1">int</span><span class="s3">>> graph;</span></p><p class="p1"><span class="s2">    </span><span class="s1">for</span><span class="s2">(</span><span class="s1">int</span><span class="s2"> i = </span><span class="s4">0</span><span class="s2">; i < prerequisites.</span><span class="s5">size</span><span class="s2">(); i++) {</span></p><p class="p1"><span class="s2">      graph[</span><span class="s5">prerequisites</span><span class="s2">[</span><span class="s5">i</span><span class="s2">].</span><span class="s5">first</span><span class="s2">].</span><span class="s5">push_back</span><span class="s2">(prerequisites[</span><span class="s5">i</span><span class="s2">].</span><span class="s5">second</span><span class="s2">);</span></p><p class="p1"><span class="s2">    }</span></p><p class="p3"><span class="s2">    </span></p><p class="p1"><span class="s2">    </span><span class="s1">for</span><span class="s2"> (</span><span class="s1">int</span><span class="s2"> i = </span><span class="s4">0</span><span class="s2">; i < numCourses; ++i)</span></p><p class="p1"><span class="s2">    {</span></p><p class="p1"><span class="s2">      </span><span class="s1">if</span><span class="s2"> (record[i] == </span><span class="s6">unvisited</span><span class="s2">)</span></p><p class="p1"><span class="s2">      {</span></p><p class="p1"><span class="s2">        </span><span class="s1">if</span><span class="s2"> (</span><span class="s6">top_sort</span><span class="s2">(i, graph, record, result) == </span><span class="s1">false</span><span class="s2">){</span></p><p class="p1"><span class="s2">          result.</span><span class="s5">clear</span><span class="s2">();</span></p><p class="p1"><span class="s2">          </span><span class="s1">return</span><span class="s2"> result;</span></p><p class="p1"><span class="s2">        }</span></p><p class="p1"><span class="s2">      }</span></p><p class="p1"><span class="s2">    }</span></p><p class="p1"><span class="s2">    </span><span class="s1">for</span><span class="s2">(</span><span class="s1">int</span><span class="s2"> i = </span><span class="s4">0</span><span class="s2">; i < result.</span><span class="s5">size</span><span class="s2">(); i++)</span></p><p class="p1"><span class="s2">      </span><span class="s5">cout</span><span class="s2"> << result[</span><span class="s5">i</span><span class="s2">] << </span><span class="s5">endl</span><span class="s2">;</span></p><p class="p1"><span class="s2">    </span><span class="s1">return</span><span class="s2"> result;</span></p><p class="p1"><span class="s2">  }</span></p><p class="p3"><span class="s2">  </span></p><p class="p2"><span class="s2">private</span><span class="s3">:</span></p><p class="p1"><span class="s2">  </span><span class="s1">bool</span><span class="s2"> top_sort(</span><span class="s1">int</span><span class="s2"> target, </span><span class="s5">unordered_map</span><span class="s2"><</span><span class="s1">int</span><span class="s2">, </span><span class="s5">vector</span><span class="s2"><</span><span class="s1">int</span><span class="s2">>> &graph, </span><span class="s6">Status</span><span class="s2"> *record, </span><span class="s5">vector</span><span class="s2"><</span><span class="s1">int</span><span class="s2">> &result) {</span></p><p class="p5"><span class="s3">    </span><span class="s2">//return if we have visited</span></p><p class="p1"><span class="s2">    </span><span class="s1">if</span><span class="s2"> (record[target] == </span><span class="s6">visited</span><span class="s2">)</span></p><p class="p2"><span class="s3">      </span><span class="s2">return</span><span class="s3"> </span><span class="s2">true</span><span class="s3">;</span></p><p class="p5"><span class="s3">    </span><span class="s2">//return if we have reached the end of the graph</span></p><p class="p1"><span class="s2">    </span><span class="s1">else</span><span class="s2"> </span><span class="s1">if</span><span class="s2"> (graph.</span><span class="s5">find</span><span class="s2">(target) == graph.</span><span class="s5">end</span><span class="s2">())</span></p><p class="p1"><span class="s2">    {</span></p><p class="p1"><span class="s2">      record[target] = </span><span class="s6">visited</span><span class="s2">;</span></p><p class="p1"><span class="s2">      result.</span><span class="s5">push_back</span><span class="s2">(target);</span></p><p class="p2"><span class="s3">      </span><span class="s2">return</span><span class="s3"> </span><span class="s2">true</span><span class="s3">;</span></p><p class="p1"><span class="s2">    }</span></p><p class="p5"><span class="s3">    </span><span class="s2">//there is a cycle in the graph</span></p><p class="p1"><span class="s2">    </span><span class="s1">else</span><span class="s2"> </span><span class="s1">if</span><span class="s2"> (record[target] == </span><span class="s6">temp_visited</span><span class="s2">)</span></p><p class="p2"><span class="s3">      </span><span class="s2">return</span><span class="s3"> </span><span class="s2">false</span><span class="s3">;</span></p><p class="p1"><span class="s2">    </span><span class="s1">else</span><span class="s2"> {</span></p><p class="p1"><span class="s2">      </span><span class="s5">vector</span><span class="s2"><</span><span class="s1">int</span><span class="s2">> adjacency_node = graph[</span><span class="s5">target</span><span class="s2">];</span></p><p class="p1"><span class="s2">      record[target] = </span><span class="s6">temp_visited</span><span class="s2">;</span></p><p class="p1"><span class="s2">      </span><span class="s1">for</span><span class="s2"> (</span><span class="s1">int</span><span class="s2"> i = </span><span class="s4">0</span><span class="s2">; i < adjacency_node.</span><span class="s5">size</span><span class="s2">(); ++i)</span></p><p class="p1"><span class="s2">      {</span></p><p class="p1"><span class="s2">    <span>	</span><span>	</span><span>	</span><span>	</span></span><span class="s1">if</span><span class="s2"> (</span><span class="s6">top_sort</span><span class="s2">(adjacency_node[</span><span class="s5">i</span><span class="s2">],graph,record,result) == </span><span class="s1">false</span><span class="s2">)</span></p><p class="p1"><span class="s2">              </span><span class="s1">return</span><span class="s2"> </span><span class="s1">false</span><span class="s2">;</span></p><p class="p1"><span class="s2">      }</span></p><p class="p1"><span class="s2">      result.</span><span class="s5">push_back</span><span class="s2">(target);</span></p><p class="p1"><span class="s2">      record[target] = </span><span class="s6">visited</span><span class="s2">;</span></p><p class="p2"><span class="s3">      </span><span class="s2">return</span><span class="s3"> </span><span class="s2">true</span><span class="s3">;</span></p><p class="p1"><span class="s2">    }</span></p><p class="p1"><span class="s2">  }</span></p><p class="p1"><span class="s2">};</span></p>

需要注意的一点是必须先递归再push进vector，而不能先push再递归。

如果我们每次都是从图的leaf开始排序，leaf每次都是先入vector。所以我们需要保证consistency，即无论以哪个node为始排序，皆得保证靠近leaf的node先入vector。所以得先递归再oush进vector。

若要改变result输出顺序，再构建图时，将图的方向颠倒即可。