Skip to content

strsplit on comma, with a trailing comma in input

5 messages · Dan Kelley, Benilton Carvalho, Gabor Grothendieck +1 more

Original
Dan Kelley
# 
I have a comma-separated data file in which trailing commas sometimes occur. 
I am using strsplit to extract the data from this file, and it seems great
except in cases with trailing comma characters.

The example below illustrates.  What I'd like is to get a fourth element in
the answer, being an empty string just like the second element.  Is there a
way I can express my patter (or perhaps specify perl or extended) to get
that?
[1] "a" ""  "b"
Benilton Carvalho
# 
my understanding is that this behavior is known (the help file  
mentions something along these lines in the example).

i'd use something like:

theText <- "a,,b,"
theText <- gsub("\\,$", ", ", theText)

and then use strsplit() on "theText"

b
On Dec 3, 2007, at 6:22 PM, dankelley wrote:

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Gabor Grothendieck
    

    

      
      #
    

  


  

      
Try appending another comma:

strsplit(paste("a,,b,", ",", sep = ""), ",")

        
On Dec 3, 2007 6:22 PM, dankelley <Dan.Kelley at dal.ca> wrote:

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Peter Dalgaard
    

    

      
      #
    

  


  

      
dankelley wrote:

            

      
      
      
    

        
Hmm, I don't think strsplit can do that. However:

 > scan(textConnection("a,,b,"), sep=",", what="")
Read 4 items
[1] "a" ""  "b" ""

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Dan Kelley
    

    

      
      #
    

  


  

      
This works perfectly.  Thanks!

        
Peter Dalgaard wrote: