Skip to content

Remove cases with -Inf from a data frame

8 messages · AbouEl-Makarim Aboueissa, Michael Dewey, Martin Møller Skarbiniks Pedersen +3 more

AbouEl-Makarim Aboueissa
# 
Dear All: good morning


I have a log-transformed data frame with some *-Inf* data values.

*my question: *how to remove all rows with *-Inf* data value from that data
frame?


with many thanks
abou
______________________


*AbouEl-Makarim Aboueissa, PhD*

*Professor, Statistics and Data Science*
*Graduate Coordinator*

*Department of Mathematics and Statistics*
*University of Southern Maine*
Michael Dewey
# 
Dear Abou

Depends on exact details of your variables but

?is.finite

Gives you the basic tool.
On 16/02/2019 15:05, AbouEl-Makarim Aboueissa wrote:

  
    
  


  


      

    

    

      
      

        


  

        

      Martin Møller Skarbiniks Pedersen
    

    

      
      #
    

  


  

      
On Sat, 16 Feb 2019 at 16:07, AbouEl-Makarim Aboueissa <

        
abouelmakarim1962 at gmail.com> wrote:

            

      
      
      
    

        
data

            

      
      
      
    

        
Hi,
  Here is a solution which uses apply.

First a data-frame as input:

set.seed(1)
df <- data.frame(w = sample(c(-Inf,1:20), 10),
                 x = sample(c(-Inf,1:20), 10),
                 y = sample(c(-Inf,1:20), 10),
                 z = sample(c(-Inf,1:20), 10))

df <- df[-(unlist(apply(df, 2, function(x) which(x == -Inf)))),]

Regards
Martin

  
  


  


      

    

    

      
      

        


  

        

      Bert Gunter
    

    

      
      #
    

  


  

      
Many ways. I assume you know that Inf and -Inf are (special) numeric values
that can be treated like other numerics. i.e.

            

      
      
      
    

        
[1] FALSE

So straightforward indexing (selection) would do it.
But there is also ?is.infinite and ?is.finite, so

apply(yourdat, 1, function(x)all(is.finite(x)))

would produce the index vector to keep rows with only finite values
assuming yourdat contains only numeric data. If this is not the case, just
select the numeric columns to index on, i.e.

apply(yourdat[sapply(yourdat,is.numeric)], 1, function(x) all(is.finite(x)))

One possible problem here is handling of NA's:

is.finite(c(-Inf,NA))
[1] FALSE FALSE
... so rows containing NA's but no -Inf's would also get removed. If you
wish to keep rows with NA's but no -Inf's, then

function(x)(is.finite(x) | is.na(x) )

could be used.

Cheers,
Bert

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 Sat, Feb 16, 2019 at 7:07 AM AbouEl-Makarim Aboueissa <

        
abouelmakarim1962 at gmail.com> wrote:

        

            

      
      
      
    

        

      
      
    

        

  
  


  


      

    

    

      
      

        


  

        

      Bert Gunter
    

    

      
      #
    

  


  

      
Sorry, that's

function(x)all(is.finite(x) | is.na(x) )

of course.


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 Sat, Feb 16, 2019 at 8:25 AM Bert Gunter <bgunter.4567 at gmail.com> wrote:

        

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

  
  


  


      

    

    

      
      

        


  

        

      Rui Barradas
    

    

      
      #
    

  


  

      
Hello,

An alternative, same dataset.

df[apply(df, 1, function(x) all(is.finite(x))), ]


Hope this helps,

Rui Barradas

?s 16:14 de 16/02/2019, Martin M?ller Skarbiniks Pedersen escreveu:

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      AbouEl-Makarim Aboueissa
    

    

      
      #
    

  


  

      
Dear Rui and All:

thank you very much for your very helpful responses.

with many thanks
abou
______________________


*AbouEl-Makarim Aboueissa, PhD*

*Professor, Statistics and Data Science*
*Graduate Coordinator*

*Department of Mathematics and Statistics*
*University of Southern Maine*

        
On Sat, Feb 16, 2019 at 11:36 AM Rui Barradas <ruipbarradas at sapo.pt> wrote:

        

            

      
      
      
    

        

      
      
    

        

      
      
    

        

  
  


  


      

    

    

      
      

        


  

        

      Ek Esawi
    

    

      
      #
    

  


  

      
This is a similar versions of other answers.

df[apply(apply(df,2,is.finite),1,sum)==4,]

BOL---EK

On Sat, Feb 16, 2019 at 10:07 AM AbouEl-Makarim Aboueissa

        
<abouelmakarim1962 at gmail.com> wrote: