nl.topicus.jdbc.shaded.com.google.common.math.Stats Maven / Gradle / Ivy
Show all versions of spanner-jdbc Show documentation
/*
* Copyright (C) 2012 The Guava Authors
*
* Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except
* in compliance with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software distributed under the License
* is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express
* or implied. See the License for the specific language governing permissions and limitations under
* the License.
*/
package nl.topicus.jdbc.shaded.com.google.common.math;
import static nl.topicus.jdbc.shaded.com.google.common.base.Preconditions.checkArgument;
import static nl.topicus.jdbc.shaded.com.google.common.base.Preconditions.checkNotNull;
import static nl.topicus.jdbc.shaded.com.google.common.base.Preconditions.checkState;
import static nl.topicus.jdbc.shaded.com.google.common.math.DoubleUtils.ensureNonNegative;
import static nl.topicus.jdbc.shaded.com.google.common.math.StatsAccumulator.calculateNewMeanNonFinite;
import static nl.topicus.jdbc.shaded.com.google.common.primitives.Doubles.isFinite;
import static java.lang.Double.NaN;
import static java.lang.Double.doubleToLongBits;
import static java.lang.Double.isNaN;
import nl.topicus.jdbc.shaded.com.google.common.annotations.Beta;
import nl.topicus.jdbc.shaded.com.google.common.annotations.GwtIncompatible;
import nl.topicus.jdbc.shaded.com.google.common.base.MoreObjects;
import nl.topicus.jdbc.shaded.com.google.common.base.Objects;
import java.io.Serializable;
import java.nio.ByteBuffer;
import java.nio.ByteOrder;
import java.util.Iterator;
import nl.topicus.jdbc.shaded.javax.annotation.Nullable;
/**
* A bundle of statistical summary values -- sum, count, mean/average, min and max, and several
* forms of variance -- that were computed from a single set of zero or more floating-point values.
*
* There are two ways to obtain a {@code Stats} instance:
*
*
* - If all the values you want to summarize are already known, use the appropriate {@code
* Stats.of} factory method below. Primitive arrays, iterables and iterators of any kind of
* {@code Number}, and primitive varargs are supported.
*
- Or, to avoid storing up all the data first, create a {@link StatsAccumulator} instance, feed
* values to it as you get them, then call {@link StatsAccumulator#snapshot}.
*
*
* Static convenience methods called {@code meanOf} are also provided for users who wish to
* calculate only the mean.
*
*
Java 8 users: If you are not using any of the variance statistics, you may wish to use
* built-in JDK libraries instead of this class.
*
* @author Pete Gillin
* @author Kevin Bourrillion
* @since 20.0
*/
@Beta
@GwtIncompatible
public final class Stats implements Serializable {
private final long count;
private final double mean;
private final double sumOfSquaresOfDeltas;
private final double min;
private final double max;
/**
* Internal constructor. Users should use {@link #of} or {@link StatsAccumulator#snapshot}.
*
*
To ensure that the created instance obeys its contract, the parameters should satisfy the
* following constraints. This is the callers responsibility and is not enforced here.
*
* - If {@code count} is 0, {@code mean} may have any finite value (its only usage will be to
* get multiplied by 0 to calculate the sum), and the other parameters may have any values (they
* will not be used).
*
- If {@code count} is 1, {@code sumOfSquaresOfDeltas} must be exactly 0.0 or
* {@link Double#NaN}.
*
*/
Stats(long count, double mean, double sumOfSquaresOfDeltas, double min, double max) {
this.count = count;
this.mean = mean;
this.sumOfSquaresOfDeltas = sumOfSquaresOfDeltas;
this.min = min;
this.max = max;
}
/**
* Returns statistics over a dataset containing the given values.
*
* @param values a series of values, which will be converted to {@code double} values (this may
* cause loss of precision)
*/
public static Stats of(Iterable values) {
StatsAccumulator accumulator = new StatsAccumulator();
accumulator.addAll(values);
return accumulator.snapshot();
}
/**
* Returns statistics over a dataset containing the given values.
*
* @param values a series of values, which will be converted to {@code double} values (this may
* cause loss of precision)
*/
public static Stats of(Iterator values) {
StatsAccumulator accumulator = new StatsAccumulator();
accumulator.addAll(values);
return accumulator.snapshot();
}
/**
* Returns statistics over a dataset containing the given values.
*
* @param values a series of values
*/
public static Stats of(double... values) {
StatsAccumulator acummulator = new StatsAccumulator();
acummulator.addAll(values);
return acummulator.snapshot();
}
/**
* Returns statistics over a dataset containing the given values.
*
* @param values a series of values
*/
public static Stats of(int... values) {
StatsAccumulator acummulator = new StatsAccumulator();
acummulator.addAll(values);
return acummulator.snapshot();
}
/**
* Returns statistics over a dataset containing the given values.
*
* @param values a series of values, which will be converted to {@code double} values (this may
* cause loss of precision for longs of magnitude over 2^53 (slightly over 9e15))
*/
public static Stats of(long... values) {
StatsAccumulator acummulator = new StatsAccumulator();
acummulator.addAll(values);
return acummulator.snapshot();
}
/**
* Returns the number of values.
*/
public long count() {
return count;
}
/**
* Returns the arithmetic mean of the
* values. The count must be non-zero.
*
* If these values are a sample drawn from a population, this is also an unbiased estimator of
* the arithmetic mean of the population.
*
*
Non-finite values
*
* If the dataset contains {@link Double#NaN} then the result is {@link Double#NaN}. If it
* contains both {@link Double#POSITIVE_INFINITY} and {@link Double#NEGATIVE_INFINITY} then the
* result is {@link Double#NaN}. If it contains {@link Double#POSITIVE_INFINITY} and finite values
* only or {@link Double#POSITIVE_INFINITY} only, the result is {@link Double#POSITIVE_INFINITY}.
* If it contains {@link Double#NEGATIVE_INFINITY} and finite values only or
* {@link Double#NEGATIVE_INFINITY} only, the result is {@link Double#NEGATIVE_INFINITY}.
*
*
If you only want to calculate the mean, use {#meanOf} instead of creating a {@link Stats}
* instance.
*
* @throws IllegalStateException if the dataset is empty
*/
public double mean() {
checkState(count != 0);
return mean;
}
/**
* Returns the sum of the values.
*
*
Non-finite values
*
* If the dataset contains {@link Double#NaN} then the result is {@link Double#NaN}. If it
* contains both {@link Double#POSITIVE_INFINITY} and {@link Double#NEGATIVE_INFINITY} then the
* result is {@link Double#NaN}. If it contains {@link Double#POSITIVE_INFINITY} and finite values
* only or {@link Double#POSITIVE_INFINITY} only, the result is {@link Double#POSITIVE_INFINITY}.
* If it contains {@link Double#NEGATIVE_INFINITY} and finite values only or
* {@link Double#NEGATIVE_INFINITY} only, the result is {@link Double#NEGATIVE_INFINITY}.
*/
public double sum() {
return mean * count;
}
/**
* Returns the population
* variance of the values. The count must be non-zero.
*
*
This is guaranteed to return zero if the dataset contains only exactly one finite value.
* It is not guaranteed to return zero when the dataset consists of the same value multiple times,
* due to numerical errors. However, it is guaranteed never to return a negative result.
*
*
Non-finite values
*
* If the dataset contains any non-finite values ({@link Double#POSITIVE_INFINITY},
* {@link Double#NEGATIVE_INFINITY}, or {@link Double#NaN}) then the result is {@link Double#NaN}.
*
* @throws IllegalStateException if the dataset is empty
*/
public double populationVariance() {
checkState(count > 0);
if (isNaN(sumOfSquaresOfDeltas)) {
return NaN;
}
if (count == 1) {
return 0.0;
}
return ensureNonNegative(sumOfSquaresOfDeltas) / count();
}
/**
* Returns the
*
* population standard deviation of the values. The count must be non-zero.
*
*
This is guaranteed to return zero if the dataset contains only exactly one finite value.
* It is not guaranteed to return zero when the dataset consists of the same value multiple times,
* due to numerical errors. However, it is guaranteed never to return a negative result.
*
*
Non-finite values
*
* If the dataset contains any non-finite values ({@link Double#POSITIVE_INFINITY},
* {@link Double#NEGATIVE_INFINITY}, or {@link Double#NaN}) then the result is {@link Double#NaN}.
*
* @throws IllegalStateException if the dataset is empty
*/
public double populationStandardDeviation() {
return Math.sqrt(populationVariance());
}
/**
* Returns the unbaised sample
* variance of the values. If this dataset is a sample drawn from a population, this is an
* unbiased estimator of the population variance of the population. The count must be greater than
* one.
*
*
This is not guaranteed to return zero when the dataset consists of the same value multiple
* times, due to numerical errors. However, it is guaranteed never to return a negative result.
*
*
Non-finite values
*
* If the dataset contains any non-finite values ({@link Double#POSITIVE_INFINITY},
* {@link Double#NEGATIVE_INFINITY}, or {@link Double#NaN}) then the result is {@link Double#NaN}.
*
* @throws IllegalStateException if the dataset is empty or contains a single value
*/
public double sampleVariance() {
checkState(count > 1);
if (isNaN(sumOfSquaresOfDeltas)) {
return NaN;
}
return ensureNonNegative(sumOfSquaresOfDeltas) / (count - 1);
}
/**
* Returns the
*
* corrected sample standard deviation of the values. If this dataset is a sample drawn from a
* population, this is an estimator of the population standard deviation of the population which
* is less biased than {@link #populationStandardDeviation()} (the unbiased estimator depends on
* the distribution). The count must be greater than one.
*
*
This is not guaranteed to return zero when the dataset consists of the same value multiple
* times, due to numerical errors. However, it is guaranteed never to return a negative result.
*
*
Non-finite values
*
* If the dataset contains any non-finite values ({@link Double#POSITIVE_INFINITY},
* {@link Double#NEGATIVE_INFINITY}, or {@link Double#NaN}) then the result is {@link Double#NaN}.
*
* @throws IllegalStateException if the dataset is empty or contains a single value
*/
public double sampleStandardDeviation() {
return Math.sqrt(sampleVariance());
}
/**
* Returns the lowest value in the dataset. The count must be non-zero.
*
*
Non-finite values
*
* If the dataset contains {@link Double#NaN} then the result is {@link Double#NaN}. If it
* contains {@link Double#NEGATIVE_INFINITY} and not {@link Double#NaN} then the result is
* {@link Double#NEGATIVE_INFINITY}. If it contains {@link Double#POSITIVE_INFINITY} and finite
* values only then the result is the lowest finite value. If it contains
* {@link Double#POSITIVE_INFINITY} only then the result is {@link Double#POSITIVE_INFINITY}.
*
* @throws IllegalStateException if the dataset is empty
*/
public double min() {
checkState(count != 0);
return min;
}
/**
* Returns the highest value in the dataset. The count must be non-zero.
*
*
Non-finite values
*
* If the dataset contains {@link Double#NaN} then the result is {@link Double#NaN}. If it
* contains {@link Double#POSITIVE_INFINITY} and not {@link Double#NaN} then the result is
* {@link Double#POSITIVE_INFINITY}. If it contains {@link Double#NEGATIVE_INFINITY} and finite
* values only then the result is the highest finite value. If it contains
* {@link Double#NEGATIVE_INFINITY} only then the result is {@link Double#NEGATIVE_INFINITY}.
*
* @throws IllegalStateException if the dataset is empty
*/
public double max() {
checkState(count != 0);
return max;
}
/**
* {@inheritDoc}
*
*
Note: This tests exact equality of the calculated statistics, including the floating
* point values. It is definitely true for instances constructed from exactly the same values in
* the same order. It is also true for an instance round-tripped through java serialization.
* However, floating point rounding errors mean that it may be false for some instances where the
* statistics are mathematically equal, including the same values in a different order.
*/
@Override
public boolean equals(@Nullable Object obj) {
if (obj == null) {
return false;
}
if (getClass() != obj.getClass()) {
return false;
}
Stats other = (Stats) obj;
return (count == other.count)
&& (doubleToLongBits(mean) == doubleToLongBits(other.mean))
&& (doubleToLongBits(sumOfSquaresOfDeltas) == doubleToLongBits(other.sumOfSquaresOfDeltas))
&& (doubleToLongBits(min) == doubleToLongBits(other.min))
&& (doubleToLongBits(max) == doubleToLongBits(other.max));
}
/**
* {@inheritDoc}
*
*
Note: This hash code is consistent with exact equality of the calculated statistics,
* including the floating point values. See the note on {@link #equals} for details.
*/
@Override
public int hashCode() {
return Objects.hashCode(count, mean, sumOfSquaresOfDeltas, min, max);
}
@Override
public String toString() {
if (count() > 0) {
return MoreObjects.toStringHelper(this)
.add("count", count)
.add("mean", mean)
.add("populationStandardDeviation", populationStandardDeviation())
.add("min", min)
.add("max", max)
.toString();
} else {
return MoreObjects.toStringHelper(this).add("count", count).toString();
}
}
double sumOfSquaresOfDeltas() {
return sumOfSquaresOfDeltas;
}
/**
* Returns the arithmetic mean of the
* values. The count must be non-zero.
*
*
The definition of the mean is the same as {@link Stats#mean}.
*
* @param values a series of values, which will be converted to {@code double} values (this may
* cause loss of precision)
* @throws IllegalArgumentException if the dataset is empty
*/
public static double meanOf(Iterable values) {
return meanOf(values.iterator());
}
/**
* Returns the arithmetic mean of the
* values. The count must be non-zero.
*
*
The definition of the mean is the same as {@link Stats#mean}.
*
* @param values a series of values, which will be converted to {@code double} values (this may
* cause loss of precision)
* @throws IllegalArgumentException if the dataset is empty
*/
public static double meanOf(Iterator values) {
checkArgument(values.hasNext());
long count = 1;
double mean = values.next().doubleValue();
while (values.hasNext()) {
double value = values.next().doubleValue();
count++;
if (isFinite(value) && isFinite(mean)) {
// Art of Computer Programming vol. 2, Knuth, 4.2.2, (15)
mean += (value - mean) / count;
} else {
mean = calculateNewMeanNonFinite(mean, value);
}
}
return mean;
}
/**
* Returns the arithmetic mean of the
* values. The count must be non-zero.
*
*
The definition of the mean is the same as {@link Stats#mean}.
*
* @param values a series of values
* @throws IllegalArgumentException if the dataset is empty
*/
public static double meanOf(double... values) {
checkArgument(values.length > 0);
double mean = values[0];
for (int index = 1; index < values.length; index++) {
double value = values[index];
if (isFinite(value) && isFinite(mean)) {
// Art of Computer Programming vol. 2, Knuth, 4.2.2, (15)
mean += (value - mean) / (index + 1);
} else {
mean = calculateNewMeanNonFinite(mean, value);
}
}
return mean;
}
/**
* Returns the arithmetic mean of the
* values. The count must be non-zero.
*
*
The definition of the mean is the same as {@link Stats#mean}.
*
* @param values a series of values
* @throws IllegalArgumentException if the dataset is empty
*/
public static double meanOf(int... values) {
checkArgument(values.length > 0);
double mean = values[0];
for (int index = 1; index < values.length; index++) {
double value = values[index];
if (isFinite(value) && isFinite(mean)) {
// Art of Computer Programming vol. 2, Knuth, 4.2.2, (15)
mean += (value - mean) / (index + 1);
} else {
mean = calculateNewMeanNonFinite(mean, value);
}
}
return mean;
}
/**
* Returns the arithmetic mean of the
* values. The count must be non-zero.
*
*
The definition of the mean is the same as {@link Stats#mean}.
*
* @param values a series of values, which will be converted to {@code double} values (this may
* cause loss of precision for longs of magnitude over 2^53 (slightly over 9e15))
* @throws IllegalArgumentException if the dataset is empty
*/
public static double meanOf(long... values) {
checkArgument(values.length > 0);
double mean = values[0];
for (int index = 1; index < values.length; index++) {
double value = values[index];
if (isFinite(value) && isFinite(mean)) {
// Art of Computer Programming vol. 2, Knuth, 4.2.2, (15)
mean += (value - mean) / (index + 1);
} else {
mean = calculateNewMeanNonFinite(mean, value);
}
}
return mean;
}
// Serialization helpers
/**
* The size of byte array representaion in bytes.
*/
static final int BYTES = (Long.SIZE + Double.SIZE * 4) / Byte.SIZE;
/**
* Gets a byte array representation of this instance.
*
*
Note: No guarantees are made regarding stability of the representation between
* versions.
*/
public byte[] toByteArray() {
ByteBuffer buff = ByteBuffer.allocate(BYTES).order(ByteOrder.LITTLE_ENDIAN);
writeTo(buff);
return buff.array();
}
/**
* Writes to the given {@link ByteBuffer} a byte representation of this instance.
*
*
Note: No guarantees are made regarding stability of the representation between
* versions.
*
* @param buffer A {@link ByteBuffer} with at least BYTES {@link ByteBuffer#remaining}, ordered as
* {@link ByteOrder#LITTLE_ENDIAN}, to which a BYTES-long byte representation of this instance
* is written. In the process increases the position of {@link ByteBuffer} by BYTES.
*/
void writeTo(ByteBuffer buffer) {
checkNotNull(buffer);
checkArgument(
buffer.remaining() >= BYTES,
"Expected at least Stats.BYTES = %s remaining , got %s",
BYTES,
buffer.remaining());
buffer
.putLong(count)
.putDouble(mean)
.putDouble(sumOfSquaresOfDeltas)
.putDouble(min)
.putDouble(max);
}
/**
* Creates a Stats instance from the given byte representation which was obtained by
* {@link #toByteArray}.
*
*
Note: No guarantees are made regarding stability of the representation between
* versions.
*/
public static Stats fromByteArray(byte[] byteArray) {
checkNotNull(byteArray);
checkArgument(
byteArray.length == BYTES,
"Expected Stats.BYTES = %s remaining , got %s",
BYTES,
byteArray.length);
return readFrom(ByteBuffer.wrap(byteArray).order(ByteOrder.LITTLE_ENDIAN));
}
/**
* Creates a Stats instance from the byte representation read from the given {@link ByteBuffer}.
*
*
Note: No guarantees are made regarding stability of the representation between
* versions.
*
* @param buffer A {@link ByteBuffer} with at least BYTES {@link ByteBuffer#remaining}, ordered as
* {@link ByteOrder#LITTLE_ENDIAN}, from which a BYTES-long byte representation of this
* instance is read. In the process increases the position of {@link ByteBuffer} by BYTES.
*/
static Stats readFrom(ByteBuffer buffer) {
checkNotNull(buffer);
checkArgument(
buffer.remaining() >= BYTES,
"Expected at least Stats.BYTES = %s remaining , got %s",
BYTES,
buffer.remaining());
return new Stats(
buffer.getLong(),
buffer.getDouble(),
buffer.getDouble(),
buffer.getDouble(),
buffer.getDouble());
}
private static final long serialVersionUID = 0;
}