cbind multiple files having same number of columns still error "arguments imply differing number of rows"

Question

I have a folder with 40 files (although I want to write the code so that it works with any number of files as this will come up all the time for me). All I want to do is obtain a new file that compiles all those files together as columns. Each file contains 5 columns, each with 902 rows.

This is what I wrote:

Specs2 <- function() 
{
  allspecs2 <- NULL
  fnames2 <- scan(file.choose(), what = "character", quiet = TRUE)
  print(fnames2)
  for (i in fnames2) 

  {
    dat2 <- read.table(paste("E:\\Post-doc Data\\Colour-   Temperature\\Compiled Sets Rows\\",i, sep=""))
    value <- data.frame(dat2)
    allspecs2 = cbind(allspecs2, value)
  }

  allspecs2 <- data.frame(allspecs2)

  write.table(allspecs2, "E:\\Post-doc Data\\Colour-   Temperature\\Colour-Temperature_Columns.csv", row.names = FALSE, quote = FALSE)
}

Specs2()

When prompted to choose file I choose a file which contains the list of all the files in the folder I want to compile. When I run this code I get the following error code:

"Error in data.frame(..., check.names = FALSE) : arguments imply differing number of rows: 0, 902"

but all the files that I am compiling have the same number of rows and columns. Does anybody know why I'm getting this error or how to fix it? or a different way of coding it to achieve the same thing?

Thank you!

EDIT: So after reading MrFlicks answer, I understand why his code works but I've used the same type of code before with success:

Specs1 <- function() 
{
  allspecs1 <- NULL
  dirnames <- scan(file.choose(), what = "character", quiet = TRUE)
  print(dirnames)
  for (j in dirnames) 

{ 
 fnames1 <- list.files(path = paste("E:\\Post-doc Data\\Colour-Temperature\\Specs\\",j, sep = ""), pattern = NULL, all.files = FALSE, 
                      full.names = FALSE, recursive = FALSE,
                      ignore.case = TRUE, include.dirs = FALSE, no.. = FALSE)
print(fnames1)
for (i in fnames1) 

    {
      dat1 <- read.table(paste("E:\\Post-doc Data\\Colour-Temperature\\Specs\\",j,"\\",i, sep=""), skip = 17)
      smoothed <- ksmooth(dat1[,1], dat1[,2], kernel = "box", bandwidth = 10, x.points = seq(300, 2100, 2))
      allspecs1 = cbind(allspecs1, smoothed$y)          
    }

    allspecs1 = cbind(smoothed$x, allspecs1)
    allspecs1 <- data.frame(allspecs1)
    write.table(allspecs1, paste("E:\\Post-doc Data\\Colour-Temperature\\Compiled Sets Rows\\",j, ".txt"), row.names = FALSE,     col.names = FALSE, quote = FALSE)
    rm(allspecs1)
    allspecs1 = NULL  
  }
}

Specs1()

I just want to understand why it works some times but not others?

Thank you!

So you want a 900 row by 200 column data set? Are you sure about that? It might be nicer to have a list of 40 data frames. — Rich Scriven, Jun 06 '14 at 04:24

score 0 · Accepted Answer · answered Jun 06 '14 at 04:33

0

You get that error for the same reason that this returns that error

cbind(NULL, data.frame(a=1))
# Error in data.frame(..., check.names = FALSE) : 
#   arguments imply differing number of rows: 0, 1

You start off by definingallspecs2 <- NULL and when you go to cbind the first data.frame to that null value, you trigger the error.

It's best to check if allspecs2 has any rows/columns before cbind-ing to it. How about

Specs2 <- function() 
{
  allspecs2 <- NULL
  fnames2 <- scan(file.choose(), what = "character", quiet = TRUE)
  print(fnames2)
  for (i in fnames2) {
    value <- read.table(paste("E:\\...\\",i, sep=""))
    if ( is.null(allspecs2) ) {
        allspecs2 <- value
    } else {
        allspecs2 <- cbind(allspecs2, value)
    }
  }
  write.table(allspecs2, "E:\\...\\file.csv", row.names = FALSE, quote = FALSE)
}

Specs2()

answered Jun 06 '14 at 04:33

MrFlick

195,160
17
277
295

I've used this type of code before with success though... I don't really understand why it worked before and not this time? I've added an edit with one of the previous codes that worked. Thank you! – user3713629 Jun 06 '14 at 05:07
@user3713629 It "works" when you cbind a vector with NULL, not a data.frame. For example `cbind(1:4, NULL)` will not give the same error. In your latest code chunk `smoothed$x` is probably just a simple numeric vector. – MrFlick Jun 06 '14 at 05:13
Thank you so much! Programming is a completely new field for me and I really appreciate your help – user3713629 Jun 06 '14 at 05:16

cbind multiple files having same number of columns still error "arguments imply differing number of rows"

1 Answers1