fpCompare
Reliable Comparison of Floating Point Numbers
Comparisons of floating point numbers are problematic due to errors associated with the binary representation of decimal numbers. Despite being aware of these problems, people still use numerical methods that fail to account for these and other rounding errors (this pitfall is the first to be highlighted in Circle 1 of Burns (2012) 'The R Inferno' < http://www.burns-stat.com/pages/Tutor/R_inferno.pdf>). This package provides new relational operators useful for performing floating point number comparisons with a set tolerance.
Reliable comparison of floating point numbers
Comparisons of floating point numbers are problematic due to errors associated with the binary representation of decimal numbers. Computer scientists and programmers are aware of these problems and yet people still use numerical methods which fail to account for floating point errors (this pitfall is the first to be highlighted in Circle 1 of Burns (2012) The R Inferno).
To avoid these and other numerical rounding issues, R's help file for relational operators (e.g., ?'>') suggests using identical and all.equal when making numerical comparisons:
x1 <- 0.5 - 0.3x2 <- 0.3 - 0.1x1 == x2 # FALSE on most machinesidentical(all.equal(x1, x2), TRUE) # TRUE everywhereInspired by R FAQ 7.31 and this Stack Overflow answer, this package provides new relational operators useful for performing floating point number comparisons with a set tolerance:
fpCompare[^1] |
base |
|---|---|
%>=% |
>= |
%>>% |
> |
%<=% |
<= |
%<<% |
< |
%==% |
== |
%!=% |
!= |
These functions use the base relational operators to make comparisons, but incorporate a tolerance value (fpCompare.tolerance) similar to all.equal. The default fpCompare.tolerance value is .Machine$double.eps^0.5, set via options. This is the same default used in all.equal for numeric comparisons.
# set telorance valuetol = .Machine$double.eps^0.5 # default valueoptions(fpCompare.tolerance = tol) # perform comparisonsx1 <- 0.5 - 0.3x2 <- 0.3 - 0.1x1 == x2 # FALSE on most machinesx1 %==% x2 # TRUE everywhere[^1]: The %<<% and %>>% symbols are used instead of %<% and %>% to avoid a conflict with magrittr's pipe operator (%>%).
Installation
From CRAN
install.packages("fpCompare")From GitHub
library(devtools)install_github("PredictiveEcology/fpCompare")Bug Reports
News
Known issues: https://github.com/PredictiveEcology/fpCompare/issues
version 0.2.2
- update maintainer's email address
version 0.2.1
- update maintainer's email address
- use HTTPS for CRAN urls
- improved vignette formatting (use
rmarkdown::render)
version 0.2.0
-
add
%<<%and%>>%operators ('less than'; 'greater than') -
move
cphto end of the author field soautdisplays as first name on CRAN -
fix bug report url in documentation
-
improved documentation and formatting:
- use short urls to avoid links spilling into margins
version 0.1.0
-
initial version
-
adds new numeric relational operators that do comparisons using a tolerance value:
%>=%,%<=%,%==%,%!=%
-
use
options(fpCompare.tolerance = value)to set tolerance