ilogb: Integer truncation of logb output

In ifultools: Insightful Research Tools

Description

Calculates logb(x, base=base) + eps where eps is a user-supplied (small) value and converts the result ala as.integer. The purpose of this function is to avoid logb variations between platforms. As an example, it is conceivable that logb(32, base=2) on one platform could return a value of 5 while on another platform it could return a slightly smaller value, e.g., 4.999999999999999. The difference is due to different compiler and platform-specific implementations of the S-PLUS or R languages. If logb output is truncated (as is the case with as.integer(logb(x))) then the result may be unexpected, leading to intractable errors. By adding an appropriate and small value to logb(x), subsequent truncation results in consistent value(s) across platforms.

Usage

ilogb(x, base=2, eps=.Machine$double.eps * 1e9)

Arguments

x	a numeric or complex vector.
base	a positive or complex number: the base with respect to which logarithms are computed. Default: 2.
eps	a small numeric value that is added to the logb result prior to truncation via as.integer. Default: .Machine\$double.eps * 1e9.

Value

a numeric scalar or vector containing the result.

See Also

logScale.

Examples

## should return 3 
ilogb(8 - .Machine$double.eps, base=2)

ifultools documentation built on May 2, 2019, 4:48 p.m.