Fix an issue with HLL serialization.

HLL#fromBytes, HLL#toBytes, and HLL#clone weren't taking into account the Postgres-compatible constructor.

Author: blinsay

pom.xml

@@ -7,7 +7,7 @@
  <packaging>jar</packaging>
  <description>HyperLogLog: approximate distinct value counting algoritm</description>
  <url>https://github.com/aggregateknowledge/java-hll</url>
- <version>1.5.0</version>
+ <version>1.5.1</version>
  <name>HyperLogLog in Java</name>
  <licenses>
  <license>

src/main/java/net/agkn/hll/HLL.java

@@ -947,11 +947,21 @@ public static HLL fromBytes(final byte[] bytes) {
  final HLLType type = metadata.HLLType();
  final int regwidth = metadata.registerWidth();
  final int log2m = metadata.registerCountLog2();
- final int expthresh = metadata.log2ExplicitCutoff();
  final boolean sparseon = metadata.sparseEnabled();
 
+ final int expthresh;
+ if(metadata.explicitAuto()) {
+ expthresh = -1;
+ } else if(metadata.explicitOff()) {
+ expthresh = 0;
+ } else {
+ // NOTE: take into account that the postgres-compatible constructor
+ // subtracts one before taking a power of two.
+ expthresh = metadata.log2ExplicitCutoff() + 1;
+ }
 
  final HLL hll = new HLL(log2m, regwidth, expthresh, sparseon, type);
+
  // Short-circuit on empty, which needs no other deserialization.
  if(HLLType.EMPTY.equals(type)) {
  return hll;
@@ -1042,7 +1052,7 @@ public HLL clone() throws CloneNotSupportedException {
  //
  // Since explicitThreshold is a power of two and only has a single
  // bit set, finding the LSB is the same as finding the inverse
- copyExpthresh = BitUtil.leastSignificantBit(explicitThreshold);
+ copyExpthresh = BitUtil.leastSignificantBit(explicitThreshold) + 1;
  }
  final HLL copy = new HLL(log2m, regwidth, copyExpthresh, !sparseOff/*sparseOn*/, type);
  switch(type) {

src/test/java/net/agkn/hll/serialization/HLLSerializationTest.java

@@ -2,23 +2,19 @@
 
 import net.agkn.hll.HLL;
 import net.agkn.hll.HLLType;
-import org.testng.Assert;
 import org.testng.annotations.Test;
 
 import java.util.ArrayList;
 import java.util.Collection;
-import java.util.LinkedList;
 import java.util.List;
 import java.util.Random;
 
 import static net.agkn.hll.HLL.MAXIMUM_EXPTHRESH_PARAM;
-import static net.agkn.hll.HLL.MAXIMUM_LOG2M_PARAM;
 import static net.agkn.hll.HLL.MAXIMUM_REGWIDTH_PARAM;
 import static net.agkn.hll.HLL.MINIMUM_EXPTHRESH_PARAM;
 import static net.agkn.hll.HLL.MINIMUM_LOG2M_PARAM;
 import static net.agkn.hll.HLL.MINIMUM_REGWIDTH_PARAM;
 import static org.testng.Assert.assertEquals;
-import static org.testng.Assert.fail;
 
 /**
  * Serialization smoke-tests.
@@ -68,9 +64,13 @@ private static void assertCardinality(final HLLType hllType, final Collection<Lo
  }
  HLL copy = HLL.fromBytes(hll.toBytes());
  assertEquals(copy.cardinality(), hll.cardinality());
+ assertEquals(copy.getType(), hll.getType());
+ assertEquals(copy.toBytes(), hll.toBytes());
 
  HLL clone = hll.clone();
  assertEquals(clone.cardinality(), hll.cardinality());
+ assertEquals(clone.getType(), hll.getType());
+ assertEquals(clone.toBytes(), hll.toBytes());
  }
  }
  }