[R] Wrong bin count number with hist() ?

[ken.mailinglist at me.com]

Hi there,

# Consider the following example:

A = 19
B = 20
A< B
A==B
hist(c(1:15,B,50),breaks=c(0,15,A,50),plot=F)$count
hist(c(1:15,A,50),breaks=c(0,15,B,50),plot=F)$count

# I was expecting the same results with the following values of A and B:

A = 19.6019203953960
B = 19.6019204365543
A< B
A==B
hist(c(1:15,B,50),breaks=c(0,15,A,50),plot=F)$count# wrong
hist(c(1:15,A,50),breaks=c(0,15,B,50),plot=F)$count# correct

# Note that truehist() of MASS define the bin count number "correctly"
require(MASS)
par(mfrow=c(1,2))
     hist(c(1:15,B,50),breaks=c(0,15,A,50),col="light gray",main="wrong - hist()")
truehist(c(1:15,B,50),breaks=c(0,15,A,50),main="correct - truehist()")

# I was not able to find an explanation online or in the R documentation. 
# The option "include.lowest" doesn't help in such cases.
# Any clue to get the correct count number per bin ?

Regards,
Ken