Merge branch 'feature/random-data-sample-problem' into develop

Author: codejsha

README.md

@@ -175,8 +175,10 @@ List<String> list = Lists.asList(boxedArray);
 - Enumerate prime numbers, EPI#5.9: [c++](/cpp-algorithm/src/array/enumerate_prime_number.h) | Enumerate prime numbers in the range.
 - Order elements in an array by even and odd: [c++](/cpp-algorithm/src/array/order_element.h)(`EvenOdd`) | Order even and odd numbers in the array.
 - Order elements in an array by specified order, EPI#5.8: [c++](/cpp-algorithm/src/array/order_element.h)(`Rearrange`) | Rearrange arrays to have a specific order.
-- Random data sampling - offline, EPI#5.12: [c++](/cpp-algorithm/src/array/random_data_sampling.h)(`OfflineRandomSampling`) | Randomly select $\textit{k}$ elements from the array.
-- Random data sampling - compute permutation, EPI#5.14: [c++](/cpp-algorithm/src/array/random_data_sampling.h)(`ComputeRandomPermutation`) | Compute permutation of the array generated by random sampling.
+- Random data sampling: Select the k elements randomly from the array with uniform probability.
+  - offline, EPI#5.12: [c++](/cpp-algorithm/src/array/random_data_sampling.h)(`OfflineRandomSampling`) | Design an algorithm to return a random subset of $k$ elements from an array.
+  - online, EPI#5.13: [c++](/cpp-algorithm/src/array/random_data_sampling.h)(`OnlineRandomSampling`) | Design an algorithm that reads data and creates a random subset of size $k$.
+  - compute permutation, EPI#5.14: [c++](/cpp-algorithm/src/array/random_data_sampling.h)(`ComputeRandomPermutation`) | Compute permutation of the array generated by random sampling.
 - Replace elements
   - replace and remove: [c++](/cpp-algorithm/src/array/replace_element.h)(`ReplaceAndRemoveString1`, `ReplaceAndRemoveString2`) | Replace element and remove element in the array.
   - telex encoding: [c++](/cpp-algorithm/src/array/replace_element.h)(`TelexEncoding`) | Telex encoding for punctuation marks.

cpp-algorithm/src/array/random_data_sampling.cpp

@@ -14,6 +14,37 @@ auto RandomDataSampling::OfflineRandomSampling(const int k, std::vector<int>& ar
     return std::vector<int>{arr.begin(), arr.begin() + k};
 }
 
+auto RandomDataSampling::OnlineRandomSampling(std::vector<int>::const_iterator begin,
+                                              const std::vector<int>::const_iterator end,
+                                              const int k)
+    -> std::vector<int>
+{
+    std::vector<int> running_sample;
+    // save the first k elements
+    for (int i = 0; i < k; ++i)
+    {
+        running_sample.emplace_back(*begin++);
+    }
+
+    std::default_random_engine seed((std::random_device())());
+    int num_seen_so_far = k;
+    while (begin != end)
+    {
+        int x = *begin++;
+        ++num_seen_so_far;
+
+        // generate a random number in [0, num_seen_so_far].
+        // if the generated number exists in [0, k), replace the element with x.
+        const int idx_to_replace = std::uniform_int_distribution<int>{0, num_seen_so_far - 1}(seed);
+        if (idx_to_replace < k)
+        {
+            running_sample[idx_to_replace] = x;
+        }
+    }
+
+    return running_sample;
+}
+
 auto RandomDataSampling::ComputeRandomPermutation(const int n) -> std::vector<int>
 {
     auto permutation = std::vector<int>(n);

cpp-algorithm/src/array/random_data_sampling.h

@@ -6,23 +6,25 @@
 namespace RandomDataSampling
 {
     /**
-     * \brief Randomly select k elements from the array.
+     * \brief Select the k elements randomly from the array with uniform probability.
+     * Let A be an array with n distinct elements. Design an algorithm to return a random subset of k elements from A, where every subset has an equal chance of being chosen.
      * \param k sample size
      * \param arr input array
      * \return result array
      */
     auto OfflineRandomSampling(int k, std::vector<int>& arr) -> std::vector<int>;
 
-    // TODO: Implement OnlineRandomSampling
     /**
-     * \brief Randomly select k elements from the array.
+     * \brief Select the k elements randomly from the array with uniform probability.
+     * Design an algorithm that reads data and creates a random subset of size k, where each item has an equal chance of being included.
      * \param begin begin iterator
      * \param end end iterator
      * \param k sample size
      * \return result array
      */
-    auto OnlineRandomSampling(const std::vector<int>::const_iterator& begin,
-                              const std::vector<int>::const_iterator& end, int k)
+    auto OnlineRandomSampling(std::vector<int>::const_iterator begin,
+                              std::vector<int>::const_iterator end,
+                              int k)
         -> std::vector<int>;
 
     /**

cpp-algorithm/src/array/test/random_data_sampling_test.cpp

@@ -19,6 +19,22 @@ GTEST_TEST(RandomDataSampling, OfflineRandomSampling)
     EXPECT_EQ(k, static_cast<int>(result.size()));
 }
 
+GTEST_TEST(RandomDataSampling, OnlineRandomSampling)
+{
+    constexpr int k = 3;
+    auto arr = std::vector<int>{3, 7, 5, 11};
+    const auto result = RandomDataSampling::OnlineRandomSampling(arr.cbegin(), arr.cend(), k);
+
+    std::stringstream stream;
+    for (const int& item : result)
+    {
+        stream << item << " ";
+    }
+    std::cout << stream.str() << std::endl;
+
+    EXPECT_EQ(k, static_cast<int>(result.size()));
+}
+
 GTEST_TEST(RandomDataSampling, ComputeRandomPermutation)
 {
     constexpr int k = 3;