README.md

In MCMaurer/inclose: Making %in% close enough

R's default %in% function uses the match() function under the hood, which works similarly to identical(): only exact matches are returned. However, when working with floating point decimals, this can lead to strange results due to the limits of machine precision:

x <- seq(from = 1, to = 3, by = 0.1)
y <- c(3, 2.4, 1.8, 1.4)
x[x %in% y]
[1] 1.4 1.8 3.0

By default, R will not return 2.4 here, which can be a bit of a surprise!

Some users may be familiar with identical()'s partner in crime, all.equal(), which takes into account equality with some small tolerance, set to the square root of R's machine precision by default. However, there is no easy equivalent partner to the %in% function!

Since the default behavior is quite likely surprising to most users, the inclose package overwrites R's default %in% with an equivalent function that acts similarly to all.equal(), while rewriting R's default %in% to a function called %idin%, which stands for identical in.

Now %in% will work like this:

x <- seq(from = 1, to = 3, by = 0.1)
y <- c(3, 2.4, 1.8, 1.4)
x[x %in% y]
[1] 1.4 1.8 2.4 3.0

Thanks to the folks on this Stack Overflow thread for getting me 95% of the way there, particularly Alexis Laz.

MCMaurer/inclose documentation built on Jan. 21, 2020, 7:29 p.m.