floating point
The floating point
type holds a floating-point number, with precision dependent on the computer. It also keeps track of the IEEE floating-point exceptions raised in a calculation and an estimate of the significance with which the number should be expressed when it is converted to a string.
All arithmetic operations are carried out to the full precision of the computer; only conversions to string are affected by the estimated significance.
Version | Platforms |
---|---|
8.0.584.0 | Mac, Red Hat, SUSE, Session, Windows |
8.1.535.0 | Debian, Ubuntu |
9.0.586.0 | AIX, HP-UX, Solaris |
9.5.13.130 | Raspbian |
Creation
This is the integral over time of the number of computers reporting this property divided by the duration of the bin. It might be fractional if computers started or stopped reporting this property during the interval of the bin.
Version | Platforms |
---|---|
8.0.584.0 | Session |
Provides a measure of nonzero values, which is useful in interpreting the logarithmic results, which ignore zero values. The logarithmic results generally aren't interesting for any property that can be zero, so this inspector can be used to validate property statistics.
Version | Platforms |
---|---|
8.0.584.0 | Session |
The integral over time of the sum of all reported values, divided by the integral over time of the number of reported values. The variance, standard deviation, skewness, and kurtosis inspectors have this same domain. In particular, computers that fail and computers that report no values don't affect these statistics.
Version | Platforms |
---|---|
8.0.584.0 | Session |
Provides a measure of zero values, which is useful in interpreting the logarithmic results, which ignore zero values. The logarithmic results generally aren't interesting for any property that can be zero, so this inspector can be used to test for that issue.
Version | Platforms |
---|---|
8.0.584.0 | Session |
Properties
Returns True
if the calculation raised the divide-by-zero exception; that is, if some part of the calculation produced an exact infinity.
Version | Platforms |
---|---|
8.0.584.0 | Mac, Red Hat, SUSE, Session, Windows |
8.1.535.0 | Debian, Ubuntu |
9.0.586.0 | AIX, HP-UX, Solaris |
9.5.13.130 | Raspbian |
Returns the minimum and maximum extreme values of the list of numbers.
- Q: extrema of floating points ("3.1415926"; "-1.06"; "2.7")
- A: -1.06, 3.1415926
Version | Platforms |
---|---|
8.0.584.0 | Mac, Red Hat, SUSE, Session, Windows |
8.1.535.0 | Debian, Ubuntu |
9.0.586.0 | AIX, HP-UX, Solaris |
9.5.13.130 | Raspbian |
Returns True
if the floating point number is finite.
- Q: (it, finite of it) of (floating point "1" / 2)
- A: 0.5, True
- Q: (it, finite of it) of (floating point "1" / 0)
- A: +infinity, False
Version | Platforms |
---|---|
8.0.584.0 | Mac, Red Hat, SUSE, Session, Windows |
8.1.535.0 | Debian, Ubuntu |
9.0.586.0 | AIX, HP-UX, Solaris |
9.5.13.130 | Raspbian |
Creates a floating point type object from the specified floating point number.
- Q: floating point "3.14159265358"
- A: 3.14159265358
Version | Platforms |
---|---|
8.0.584.0 | Mac, Red Hat, SUSE, Session, Windows |
8.1.535.0 | Debian, Ubuntu |
9.0.586.0 | AIX, HP-UX, Solaris |
9.5.13.130 | Raspbian |
Returns True
if the calculation raised the inexact exception; that is, if some intermediate result could not be represented exactly.
Version | Platforms |
---|---|
8.0.584.0 | Mac, Red Hat, SUSE, Session, Windows |
8.1.535.0 | Debian, Ubuntu |
9.0.586.0 | AIX, HP-UX, Solaris |
9.5.13.130 | Raspbian |
Returns the smallest integer not less than the floating point number.
- Q: integer ceiling of floating point "3.14159265358"
- A: 4
- Q: integer ceiling of floating point "-3.14159265358"
- A: -3
Version | Platforms |
---|---|
8.0.584.0 | Mac, Red Hat, SUSE, Session, Windows |
8.1.535.0 | Debian, Ubuntu |
9.0.586.0 | AIX, HP-UX, Solaris |
9.5.13.130 | Raspbian |
Returns the largest integer less than or equal to the floating point number.
- Q: integer floor of floating point "3.14159265358"
- A: 3
- Q: integer floor of floating point "-3.14159265358"
- A: -4
Version | Platforms |
---|---|
8.0.584.0 | Mac, Red Hat, SUSE, Session, Windows |
8.1.535.0 | Debian, Ubuntu |
9.0.586.0 | AIX, HP-UX, Solaris |
9.5.13.130 | Raspbian |
Returns True
if the calculation raised the invalid exception; that is, if some part of the calculation a function was applied to a value outside its domain.
Version | Platforms |
---|---|
8.0.584.0 | Mac, Red Hat, SUSE, Session, Windows |
8.1.535.0 | Debian, Ubuntu |
9.0.586.0 | AIX, HP-UX, Solaris |
9.5.13.130 | Raspbian |
Returns True
if the value is not a number.Example: nan of (floating point "1.e-99999" * floating point "1.e999999") - Returns True
.
Version | Platforms |
---|---|
8.0.584.0 | Mac, Red Hat, SUSE, Session, Windows |
8.1.535.0 | Debian, Ubuntu |
9.0.586.0 | AIX, HP-UX, Solaris |
9.5.13.130 | Raspbian |
Returns True
if the calculation raised the overflow exception; that is, if some intermediate result was too large to be represented, but not an exact infinity.Example: overflow of (floating point "1.0e50000") - Returns True
, since the number is too big to represent in floating point.
Version | Platforms |
---|---|
8.0.584.0 | Mac, Red Hat, SUSE, Session, Windows |
8.1.535.0 | Debian, Ubuntu |
9.0.586.0 | AIX, HP-UX, Solaris |
9.5.13.130 | Raspbian |
The base 10 logarithm of the quotient of the value and its significance place; approximately the number of significant digits to which the number should be expressed.
Version | Platforms |
---|---|
8.0.584.0 | Mac, Red Hat, SUSE, Session, Windows |
8.1.535.0 | Debian, Ubuntu |
9.0.586.0 | AIX, HP-UX, Solaris |
9.5.13.130 | Raspbian |
The base 10 logarithm of the significance threshold; approximately the number of digits to the left (positive) or right (negative) of the ones place to which the number should be expressed.
Version | Platforms |
---|---|
8.0.584.0 | Mac, Red Hat, SUSE, Session, Windows |
8.1.535.0 | Debian, Ubuntu |
9.0.586.0 | AIX, HP-UX, Solaris |
9.5.13.130 | Raspbian |
The difference between the given value and the next number expressed to the same significance level. For example, the significance threshold of 3 is 1, the significance threshold of 3.0 is 0.1, and the significance threshold of 3000 is 1000.
Version | Platforms |
---|---|
8.0.584.0 | Mac, Red Hat, SUSE, Session, Windows |
8.1.535.0 | Debian, Ubuntu |
9.0.586.0 | AIX, HP-UX, Solaris |
9.5.13.130 | Raspbian |
Returns True
if the calculation raised the underflow exception; that is, if some intermediate result was a nonzero value too small to be represented.
Version | Platforms |
---|---|
8.0.584.0 | Mac, Red Hat, SUSE, Session, Windows |
8.1.535.0 | Debian, Ubuntu |
9.0.586.0 | AIX, HP-UX, Solaris |
9.5.13.130 | Raspbian |
Casts
This casting operator is added for completeness. It takes a floating point number and casts it as a floating point number. It facilitates automatic relevance generation where the software is not aware of the input types.
Version | Platforms |
---|---|
8.0.584.0 | Mac, Red Hat, SUSE, Session, Windows |
8.1.535.0 | Debian, Ubuntu |
9.0.586.0 | AIX, HP-UX, Solaris |
9.5.13.130 | Raspbian |