Re: [R] odd behavior of numeric()

2016-04-27 Thread Jeff Newmiller
FAQ 7.31 0.2 is slightly larger than 1/5, and 1-0.8 is slightly smaller. The argument to numeric has to be an integer, so fractions are discarded. This behavior is not unique to R. -- Sent from my phone. Please excuse my brevity. On April 27, 2016 8:29:13 PM GMT+01:00, Scott Rifkin wrote: >W

Re: [R] odd behavior of numeric()

2016-04-27 Thread Richard M. Heiberger
> numeric(0.2*25) [1] 0 0 0 0 0 > numeric((1-0.8)*25) [1] 0 0 0 0 > print(c(0.2*25, (1-0.8)*25), digits=17) [1] 5. 4.9991 > 0.2*25 - (1-0.8)*25 [1] 8.881784e-16 > as.integer(c(0.2*25, (1-0.8)*25), digits=17) [1] 5 4 > ## See FAQ 7.31 for other examples of floating point

[R] odd behavior of numeric()

2016-04-27 Thread Scott Rifkin
Why does: > numeric(0.2*25) return [1] 0 0 0 0 0 but > numeric((1-0.8)*25) returns [1] 0 0 0 0 [running version 3.2.0] [Apologies if this has been asked before - it's a hard question to find specific search terms for] Thanks, Scott __ R-hel