Skip to content

alternative to subset(dataframe, select = - column) in R

7 messages · Jeff Newmiller, Rui Barradas, Luigi Marongiu +1 more

Luigi Marongiu
# 
Hello,
this is a very simple question but...
what is the vector alternative to `subset(dataframe, select = - column)`?
I tried with:
```
Error in `[.data.frame`(new, -ID) : object 'ID' not found
Error in -"ID" : invalid argument to unary operator
Error in -"ID" : invalid argument to unary operator
Error: unexpected '[' in "x = new[-["
```
Thank you
Bert Gunter
# 
x <- new[-match("ID"), names(new))]

Bert Gunter

"The trouble with having an open mind is that people keep coming along
and sticking things into it."
-- Opus (aka Berkeley Breathed in his "Bloom County" comic strip )
On Fri, Sep 24, 2021 at 7:10 AM Luigi Marongiu <marongiu.luigi at gmail.com> wrote:
Jeff Newmiller
# 
The subset function

subset( new, select = -ID )

uses non-standard evaluation... it "knows" that when you write the language symbol ID it may match to one of the column names in new, and the `-` in the select argument will be noticed before R tries to change the sign of a variable ID in your calling environment and cause subset to remove the "ID" column from the returned version of new.

The vector indexing operator does none of that, so whatever indexing arguments you specify need to be valid expressions on their own without the indexing operator.

-ID
   You don't have a variable called ID, so its sign vanity be changed

-"ID"
   Changing the sign of a character literal is not a defined operation

-["ID"]
   Standalone brackets have no meaning in R... only can be used directly adjacent to a vector to the right.

What you do have is the data frame `new`... and it has names, and among those names is "ID", so...

"ID" != names( new )
   should be a logical vector with FALSE in the position corresponding to "ID", so

new[ , "ID" != names( new ) ]

will select all of the columns other than the "ID" column.

If you want to deselect more than one column, then the %in% operator can help:

! names( new ) %in% c( "ID", "Other" )

will be a logical vector with FALSE in the corresponding positions.
On September 24, 2021 7:09:54 AM PDT, Luigi Marongiu <marongiu.luigi at gmail.com> wrote:

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Bert Gunter
    

    

      
      #
    

  


  

      
typo. Correction:

x <- new[-match("ID", names(new))]

Bert Gunter

"The trouble with having an open mind is that people keep coming along
and sticking things into it."
-- Opus (aka Berkeley Breathed in his "Bloom County" comic strip )

        
On Fri, Sep 24, 2021 at 7:31 AM Bert Gunter <bgunter.4567 at gmail.com> wrote:

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Rui Barradas
    

    

      
      #
    

  


  

      
Hello,

Like this?


mtcars[names(mtcars) != "mpg"]


Hope this helps,

Rui Barradas

?s 15:09 de 24/09/21, Luigi Marongiu escreveu:

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Luigi Marongiu
    

    

      
      #
    

  


  

      
Thank you!
```
new_out <- new[ , "ID" != names( new ) ]
new[names(new) != "ID"]
```
worked as wanted, `new[-match("ID"), names(new)]` gave the error:
`Error in match("ID") : argument "table" is missing, with no default`.
Cheers

        
On Fri, Sep 24, 2021 at 7:33 PM Rui Barradas <ruipbarradas at sapo.pt> wrote:

            

      
      
      
    

        

      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Bert Gunter
    

    

      
      #
    

  


  

      
If you correct my typo as my followup message indicated, match() works
fine. As does logical indexing with %in%. The != version will work for
your query, but does not generalize to deleting several columns. The
point is to heed Jeff N's advice.


Bert Gunter

"The trouble with having an open mind is that people keep coming along
and sticking things into it."
-- Opus (aka Berkeley Breathed in his "Bloom County" comic strip )
On Fri, Sep 24, 2021 at 12:43 PM Luigi Marongiu

        
<marongiu.luigi at gmail.com> wrote: