classIntervals: 'probs' argument to 'quantile' not allowed as documented

[AMBarbosa]

Hi,

The classIntervals help file says that 'The "quantile" style provides quantile breaks; arguments to quantile may be passed through ...', but apparently the probs argument isn't allowed, as it is already set by classIntervals:

data_censored<-c(rep(0,10), rnorm(100, mean=20,sd=1),rep(26,10))
plot(density(data_censored))
classIntervals(data_censored, style="quantile", probs = seq(0, 1, 0.25))
# Error in quantile.default(x = var, probs = seq(0, 1, 1/n), ...) : 
#  formal argument "probs" matched by multiple actual arguments

I'd suggest documenting this, or else allowing the user to choose the probs, even if their default is still seq(0, 1, 1/n) (as seen in the classIntervals code).

Regards,

[rsbivand]

Thanks for a useful report, patch coming.

[rsbivand]

By the way, n as 4 gives the expected output:

> classIntervals(data_censored, style="quantile", n=4)
style: quantile
  one of 166,650 possible partitions of this variable into 4 classes
       [0,19.22306) [19.22306,19.97196) [19.97196,20.69348)       [20.69348,26] 
                 30                  30                  30                  30 

I'll try to implement a work-around extracting the integer unique difference from the provided argument.

[rsbivand]

This is my original repo - the correct repo is https://github.com/r-spatial/classInt, and the issue there is r-spatial/classInt#49. I've committed a fix to branch q_probs there. Please check whether this resolves your problem, which in any case was using probs instead of n.