Skip to content

How to merge two functions into one?

3 messages · Jonsson, Rui Barradas

Original
Jonsson
# 
I am using the code below to calculate the correlation map between two
datasets. This code worked fine.

    dir1 <- list.files("D:thly", "*.bin", full.names = TRUE)
    dir2 <- list.files("D:002", "*.envi", full.names = TRUE)
    file_tot <- array(dim = c(1440, 720, 11, 2))
    for(i in 1:length(dir1)) {
        file_tot[, , i, 1] <- readBin(dir1[i], numeric(), size = 4, 
                                        n = 1440 * 720, signed = T)
        file_tot[, , i, 2] <- readBin(dir2[i], numeric(), size = 4, 
                                        n = 1440 * 720, signed = T)
    }
    resultscor<-apply(file_tot,c(1,2),function(x){cor(x[,1],x[,2],use =
"na.or.complete")})

I would like to calculate the correlation only when the `P-value is lower
than 0.05`. so this function bellow will do the job:

     return_cor = function(x, y)  {
        z = cor.test(x,y)
        if(z[[3]] < 0.05) {
        return(z[[5]])
          } else { 
        return(NA) 
        }
        }

However I got this error(as some pairs of my data are less then 3):

                 Error in cor.test.default(x, y) : not enough finite
observations

in order to avoid this error and return NA when there are less than 3
pairs,this function does the job:

          cor_withN <- function(...) {
          res <- try(cor.test(...)$estimate, silent=TRUE)
          ifelse(class(res)=="try-error", NA, res)
           }
Both functions worked perfectly.How can we merge both functions into one
function so we calculate correlation when P value is (certain value,
threshold) and also do the calculations even if there are less than 3 pairs.




--
View this message in context: http://r.789695.n4.nabble.com/How-to-merge-two-functions-into-one-tp4659365.html
Sent from the R help mailing list archive at Nabble.com.
Rui Barradas
# 
Hello,

Instead of try(cor.test) you could try(return_cor).
Also, isn't there a bug in return_cor? See the commented line below.

return_cor = function(x, y, tresh = 0.05)  {
         z = cor.test(x,y)
         if(z[[3]] < tresh) {
       	  return(z[[4]])  # not z[[5]], always zero
           } else {
	        return(NA)
         }
}
cor_withN <- function(...) {
           res <- try(return_cor(...), silent=TRUE)
           ifelse(class(res)=="try-error", NA, res)
}

cor_withN(1:10, 1:10 + rnorm(10))
cor_withN(1:2, 1:2 + rnorm(2))


Hope this helps,

Rui Barradas


Em 22-02-2013 15:04, Jonsson escreveu: