[Rd] Bug in cor function (PR#7689)

cig69410 at syd.odn.ne.jp cig69410 at syd.odn.ne.jp
Sun Feb 13 07:08:19 CET 2005


I can't hardly accept the result of  cor function with 
pairwize.colplete.obs or complete.obs

insert print statements in cor function,


+     if (method != "pearson") {
  +         Rank <- function(u) if (is.matrix(u))
  +             apply(u, 2, rank, na.last = "keep")
  +         else rank(u, na.last = "keep")
  +         x <- Rank(x)
  +         print(x) # add
  +         if (!is.null(y)) {
  +             y <- Rank(y)
  +             print(y) # add
  +         }
  +     }
  +     .Internal(cor(x, y, na.method, method == "kendall"))

and, data is
  > x <- c(7, 9, 8,  0, NA, NA)
  > y <- c(2, 3, 4, NA,  4,  3)

and, call cor function
  > cor(x, y, use="pair", method="sp")

order of x, and y are
  [1]  2  4  3  1 NA NA
  [1] 1.0 2.5 4.5  NA 4.5 2.5

alas!! and the result is
  [1] 0.4271211

oh! no!!

the result must be 0.5



More information about the R-devel mailing list