public void HashedVsRandom(int realCardinality)
{
const int precision = 18;
var hllRandom = new HyperLogLog(precision);
for (var i = 0; i < realCardinality; i++)
{
var randomBytes = new byte[8];
RandomNumberGenerator.GetBytes(randomBytes);
var hashValue = BitConverter.ToUInt64(randomBytes, 0);
hllRandom.Add(hashValue);
}
var cardinalityEstimateRandom = hllRandom.Count();
var hllHashed = new HyperLogLog(precision);
using (var hashAlgorithm = new SHA1CryptoServiceProvider())
{
for (var i = 0; i < realCardinality; i++)
{
var randomBytes = new byte[12];
RandomNumberGenerator.GetBytes(randomBytes);
var hashValue = HyperLogLog.Hash(hashAlgorithm, randomBytes);
hllHashed.Add(hashValue);
}
}
var cardinalityEstimateHashed = hllHashed.Count();
var percentageDifference = GetPercentageDifference(cardinalityEstimateRandom, cardinalityEstimateHashed);
Assert.True(percentageDifference < ZeroFivePercent);
}
public void HashConstantWithSHA1()
{
const ulong value = 0x2D51AF5C52FDE6B4ul;
using (var hashAlgorithm = new SHA1CryptoServiceProvider())
{
var hashValue = HyperLogLog.Hash(hashAlgorithm, BitConverter.GetBytes(value));
Assert.Equal(17851087020509344997ul, hashValue);
}
}
public void HashConstantWithMD5()
{
const ulong value = 0x2D51AF5C52FDE6B4ul;
using (var hashAlgorithm = new MD5CryptoServiceProvider())
{
var hashValue = HyperLogLog.Hash(hashAlgorithm, BitConverter.GetBytes(value));
Assert.Equal(16663394367412432550ul, hashValue);
}
}
public void TheHashedRandomTheory(int realCardinality)
{
const int precision = 18;
var hll = new HyperLogLog(precision);
using (var hashAlgorithm = new SHA1CryptoServiceProvider())
{
for (var i = 0; i < realCardinality; i++)
{
var randomBytes = new byte[12];
RandomNumberGenerator.GetBytes(randomBytes);
var hashValue = HyperLogLog.Hash(hashAlgorithm, randomBytes);
hll.Add(hashValue);
}
}
var cardinalityEstimate = hll.Count();
var percentageDifference = GetPercentageDifference(realCardinality, cardinalityEstimate);
Assert.True(percentageDifference < OnePercent);
}
