*    value (whether it is NaN or infinity).
     * 2. If the new value is finite and the previous mean is non-finite then the mean is unchanged
     *    (whether it is NaN or infinity).
     * 3. If both the previous mean and the new value are non-finite and...
     * 3a. ...either or both is NaN (so mean != value) then the new mean is NaN.

 * uniform distribution on [0,1]."
 *
 * <h3>Handling of non-finite values</h3>
 *
 * <p>If any values in the input are {@link Double#NaN NaN} then all values returned are {@link
 * Double#NaN NaN}. (This is the one occasion when the behaviour is not the same as you'd get from
 * sorting with {@link java.util.Arrays#sort(double[]) Arrays.sort(double[])} or {@link

   * the arithmetic mean of the population.
   *
   * <h3>Non-finite values</h3>
   *
   * <p>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

    } else {
      // This is a generalized version of the calculation in add(double, double) above. Note that
      // non-finite inputs will have sumOfProductsOfDeltas = NaN, so non-finite values will result
      // in NaN naturally.
      sumOfProductsOfDeltas +=
          values.sumOfProductsOfDeltas()
              + (values.xStats().mean() - xStats.mean())
                  * (values.yStats().mean() - yStats.mean())

        assertFalse(DoubleMath.fuzzyEquals(a, Double.NaN, tolerance));
        assertFalse(DoubleMath.fuzzyEquals(Double.NaN, a, tolerance));
      }
    }
  }

  public void testFuzzyEqualsTwoNaNs() {
    for (double tolerance : TOLERANCE_CANDIDATES) {
      assertTrue(DoubleMath.fuzzyEquals(Double.NaN, Double.NaN, tolerance));
    }
  }

  public void testFuzzyEqualsZeroTolerance() {

          .that(Doubles.indexOf(new double[] {5.0, value, value, 5.0}, new double[] {value, value}))
          .isEqualTo(1);
    }
    assertThat(Doubles.indexOf(new double[] {5.0, NaN, NaN, 5.0}, new double[] {NaN, NaN}))
        .isEqualTo(-1);
  }

  public void testLastIndexOf() {
    assertThat(Doubles.lastIndexOf(EMPTY, 1.0)).isEqualTo(-1);
    assertThat(Doubles.lastIndexOf(ARRAY1, 2.0)).isEqualTo(-1);

        assertFalse(DoubleMath.fuzzyEquals(a, Double.NaN, tolerance));
        assertFalse(DoubleMath.fuzzyEquals(Double.NaN, a, tolerance));
      }
    }
  }

  public void testFuzzyEqualsTwoNaNs() {
    for (double tolerance : TOLERANCE_CANDIDATES) {
      assertTrue(DoubleMath.fuzzyEquals(Double.NaN, Double.NaN, tolerance));
    }
  }

  public void testFuzzyEqualsZeroTolerance() {

   *
   * <h3>Non-finite values</h3>
   *
   * <p>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 populationCovariance() {
    checkState(count() != 0);
    return sumOfProductsOfDeltas / count();
  }

    Our Id type is a prime example of this.

    >>> decimal_encoder(Decimal("1.0"))
    1.0

    >>> decimal_encoder(Decimal("1"))
    1

    >>> decimal_encoder(Decimal("NaN"))
    nan
    """
    exponent = dec_value.as_tuple().exponent
    if isinstance(exponent, int) and exponent >= 0:
        return int(dec_value)
    else:
        return float(dec_value)

        .comparingValuesUsing(QUANTILE_CORRESPONDENCE)
        .containsExactly(
            0, NaN,
            1, NaN,
            2, NaN,
            8, NaN,
            9, NaN,
            10, NaN);
  }

  public void testScale_index_compute_doubleCollection_nan() {
    assertThat(Quantiles.scale(10).index(5).compute(ONE_TO_FIVE_AND_NAN)).isNaN();
  }