Adding a unit test

src/Interpreters/BloomFilterHash.h

@@ -251,7 +251,10 @@ struct BloomFilterHash
         static const size_t MIN_BITS_PER_ROW = 2;
         static const size_t MIN_HASH_FUNCTION_COUNT = 2;
 
-        /// Return the smaller possible parameters for false positive rates bigger than 0.283
+        /// Return the smaller possible parameters for false positive rates higher or equal than 0.283
+        /// Otherwise, for those rates the loop won't find any possible values in the lookup table
+        /// returning bits_per_row = 19 & size_of_hash_functions = 13. Which are the most restrictive values
+        /// to be used with the smallest false positive rates.
         if (max_conflict_probability >= 0.283)
             return std::pair<size_t, size_t>(MIN_BITS_PER_ROW, MIN_HASH_FUNCTION_COUNT);
 

src/Interpreters/tests/gtest_calculationBestPractices.cpp

@@ -0,0 +1,57 @@
+#include <gtest/gtest.h>
+
+#include <Interpreters/BloomFilterHash.h>
+
+using namespace DB;
+
+
+TEST(BloomFilterHash, ReasonableProbabilities)
+{
+    const auto & output_1 = BloomFilterHash::calculationBestPractices(0.001);
+
+    ASSERT_EQ(output_1.first, 15);
+    ASSERT_EQ(output_1.second, 7);
+
+    const auto & output_2 = BloomFilterHash::calculationBestPractices(0.025);
+
+    ASSERT_EQ(output_2.first, 8);
+    ASSERT_EQ(output_2.second, 4);
+
+    const auto & output_3 = BloomFilterHash::calculationBestPractices(0.05);
+
+    ASSERT_EQ(output_3.first, 7);
+    ASSERT_EQ(output_3.second, 3);
+
+    const auto & output_4 = BloomFilterHash::calculationBestPractices(0.1);
+
+    ASSERT_EQ(output_4.first, 6);
+    ASSERT_EQ(output_4.second, 2);
+
+    const auto & output_5 = BloomFilterHash::calculationBestPractices(0.2);
+
+    ASSERT_EQ(output_5.first, 4);
+    ASSERT_EQ(output_5.second, 2);
+
+    const auto & output_6 = BloomFilterHash::calculationBestPractices(0.282);
+
+    ASSERT_EQ(output_6.first, 3);
+    ASSERT_EQ(output_6.second, 2);
+}
+
+TEST(BloomFilterHash, HighProbabilities)
+{
+    const auto & output_1 = BloomFilterHash::calculationBestPractices(0.283);
+
+    ASSERT_EQ(output_1.first, 2);
+    ASSERT_EQ(output_1.second, 2);
+
+    const auto & output_2 = BloomFilterHash::calculationBestPractices(0.5);
+
+    ASSERT_EQ(output_2.first, 2);
+    ASSERT_EQ(output_2.second, 2);
+
+    const auto & output_3 = BloomFilterHash::calculationBestPractices(0.8);
+
+    ASSERT_EQ(output_3.first, 2);
+    ASSERT_EQ(output_3.second, 2);
+}