From f013d71d39494fb5fe4c56a5d09213dbea16a48e Mon Sep 17 00:00:00 2001
From: Jonathan Ellis <jbellis@apache.org>
Date: Fri, 13 Aug 2010 15:29:56 +0000
Subject: [PATCH] fix errors in hard-coded bloom filter optKPerBucket by
 computing it algorithmically.  patch by Coda Hale; reviewed by jbellis for
 CASSANDRA-1220

git-svn-id: https://svn.apache.org/repos/asf/cassandra/branches/cassandra-0.6@985241 13f79535-47bb-0310-9956-ffa450edef68
---
 CHANGES.txt                                   |  2 +
 .../cassandra/utils/BloomCalculations.java    | 58 +++++++++++++------
 2 files changed, 43 insertions(+), 17 deletions(-)

diff --git a/CHANGES.txt b/CHANGES.txt
index 3fd81db3e695..14495da5eaa5 100644
--- a/CHANGES.txt
+++ b/CHANGES.txt
@@ -9,6 +9,8 @@
  * Dynamic snitch (CASSANDRA-981)
  * Use a regex to filter keyspace and column family names during
    initialization (CASSANDRA-1377)
+ * fix errors in hard-coded bloom filter optKPerBucket by computing it
+   algorithmically (CASSANDRA-1220
 
 
 0.6.4
diff --git a/src/java/org/apache/cassandra/utils/BloomCalculations.java b/src/java/org/apache/cassandra/utils/BloomCalculations.java
index 6ad6f15f748b..9d430ff83b60 100644
--- a/src/java/org/apache/cassandra/utils/BloomCalculations.java
+++ b/src/java/org/apache/cassandra/utils/BloomCalculations.java
@@ -31,10 +31,6 @@ class BloomCalculations {
 
     private static final int minBuckets = 2;
     private static final int minK = 1;
-    private static final int[] optKPerBuckets =
-            new int[]{1, // dummy K for 0 buckets per element
-                      1, // dummy K for 1 buckets per element
-                      1, 2, 3, 3, 4, 5, 5, 6, 7, 8, 8, 8, 8, 8};
 
     /**
      * In the following table, the row 'i' shows false positive rates if i buckets
@@ -47,21 +43,49 @@ class BloomCalculations {
         {1.0}, // dummy row representing 0 buckets per element
         {1.0, 1.0}, // dummy row representing 1 buckets per element
         {1.0, 0.393,  0.400},
-        {1.0, 0.283,  0.237,  0.253},
-        {1.0, 0.221,  0.155,  0.147,   0.160},
-        {1.0, 0.181,  0.109,  0.092,   0.092,   0.101}, // 5
-        {1.0, 0.154,  0.0804, 0.0609,  0.0561,  0.0578,  0.0638},
-        {1.0, 0.133,  0.0618, 0.0423,  0.0359,  0.0347,  0.0364},
-        {1.0, 0.118,  0.0489, 0.0306,  0.024,   0.0217,  0.0216,  0.0229},
-        {1.0, 0.105,  0.0397, 0.0228,  0.0166,  0.0141,  0.0133,  0.0135,  0.0145}, // 9
-        {1.0, 0.0952, 0.0329, 0.0174,  0.0118,  0.00943, 0.00844, 0.00819, 0.00846},
-        {1.0, 0.0869, 0.0276, 0.0136,  0.00864, 0.0065,  0.00552, 0.00513, 0.00509},
-        {1.0, 0.08,   0.0236, 0.0108,  0.00646, 0.00459, 0.00371, 0.00329, 0.00314},
-        {1.0, 0.074,  0.0203, 0.00875, 0.00492, 0.00332, 0.00255, 0.00217, 0.00199},
-        {1.0, 0.0689, 0.0177, 0.00718, 0.00381, 0.00244, 0.00179, 0.00146, 0.00129},
-        {1.0, 0.0645, 0.0156, 0.00596, 0.003,   0.00183, 0.00128, 0.001,   0.000852} // 15
+        {1.0, 0.283,  0.237,   0.253},
+        {1.0, 0.221,  0.155,   0.147,   0.160},
+        {1.0, 0.181,  0.109,   0.092,   0.092,   0.101}, // 5
+        {1.0, 0.154,  0.0804,  0.0609,  0.0561,  0.0578,   0.0638},
+        {1.0, 0.133,  0.0618,  0.0423,  0.0359,  0.0347,   0.0364},
+        {1.0, 0.118,  0.0489,  0.0306,  0.024,   0.0217,   0.0216,   0.0229},
+        {1.0, 0.105,  0.0397,  0.0228,  0.0166,  0.0141,   0.0133,   0.0135,   0.0145},
+        {1.0, 0.0952, 0.0329,  0.0174,  0.0118,  0.00943,  0.00844,  0.00819,  0.00846}, // 10
+        {1.0, 0.0869, 0.0276,  0.0136,  0.00864, 0.0065,   0.00552,  0.00513,  0.00509},
+        {1.0, 0.08,   0.0236,  0.0108,  0.00646, 0.00459,  0.00371,  0.00329,  0.00314},
+        {1.0, 0.074,  0.0203,  0.00875, 0.00492, 0.00332,  0.00255,  0.00217,  0.00199,  0.00194},
+        {1.0, 0.0689, 0.0177,  0.00718, 0.00381, 0.00244,  0.00179,  0.00146,  0.00129,  0.00121,  0.0012},
+        {1.0, 0.0645, 0.0156,  0.00596, 0.003,   0.00183,  0.00128,  0.001,    0.000852, 0.000775, 0.000744}, // 15
+		{1.0, 0.0606, 0.0138,  0.005,   0.00239, 0.00139,  0.000935, 0.000702, 0.000574, 0.000505, 0.00047,  0.000459},
+		{1.0, 0.0571, 0.0123,  0.00423, 0.00193, 0.00107,  0.000692, 0.000499, 0.000394, 0.000335, 0.000302, 0.000287, 0.000284},
+		{1.0, 0.054,  0.0111,  0.00362, 0.00158, 0.000839, 0.000519, 0.00036,  0.000275, 0.000226, 0.000198, 0.000183, 0.000176},
+		{1.0, 0.0513, 0.00998, 0.00312, 0.0013,  0.000663, 0.000394, 0.000264, 0.000194, 0.000155, 0.000132, 0.000118, 0.000111, 0.000109},
+		{1.0, 0.0488, 0.00906, 0.0027,  0.00108, 0.00053,  0.000303, 0.000196, 0.00014,  0.000108, 8.89e-05, 7.77e-05, 7.12e-05, 6.79e-05, 6.71e-05} // 20
     };  // the first column is a dummy column representing K=0.
 
+	/**
+	 * The optimal number of hashes for a given number of bits per element.
+	 * These values are automatically calculated from the data above.
+	 */
+	private static final int[] optKPerBuckets = new int[probs.length];
+
+    static
+    {
+        for (int i = 0; i < probs.length; i++)
+        {
+            double min = Double.MAX_VALUE;
+            double[] prob = probs[i];
+            for (int j = 0; j < prob.length; j++)
+            {
+                if (prob[j] < min)
+                {
+                    min = prob[j];
+                    optKPerBuckets[i] = Math.max(minK, j);
+                }
+            }
+        }
+    }
+
     /**
      * Given the number of buckets that can be used per element, return a
      * specification that minimizes the false positive rate.