Skip to content

double matrix?

6 messages · bruce087 at umn.edu, Don McKenzie, Richard M. Heiberger +3 more

Original
bruce087 at umn.edu
# 
Hi-

I have a 37 X 473971 character matrix that I am trying to convert into a 
numeric matrix. When I use the code:

 class(matrix) = "numeric"  

I end up with something called a "double matrix" whose dimensions are still 
37 X 473971

I have also tried 

new = apply(matrix,2, as.numeric) and got the same thing.

The analysis code I am ultimately attempting to run on this data requires 
that it be in a numerical matrix, and it is really not okay with a double 
matrix.

Does anyone know how to fix this?

Thanks.
Don McKenzie
# 
What are the entries in your matrix?  If they are something that won't coerce to numeric, you need to backtrack. Note how R distinguishes types of characters.
[1] NA
Warning message:
NAs introduced by coercion
[1] "2"
[1] 2
On Jul 31, 2013, at 1:47 PM, bruce087 at umn.edu wrote:

            

      
      
      
    

        

      
      
    

        
Don McKenzie, Research Ecologist
Pacific WIldland Fire Sciences Lab
US Forest Service

Affiliate Professor
School of Forest Resources, College of the Environment
CSES Climate Impacts Group
University of Washington

phone: 206-732-7824
dmck at uw.edu

  
  


  


      

    

    

      
      

        
      

    

    

      
      

        


  

        

      Rui Barradas
    

    

      
      #
    

  


  

      
Hello,

"double" and "numeric" are the same. From the help page for ?double, 
section "Note on names"

"It is a historical anomaly that R has two names for its floating-point 
vectors, double and numeric (and formerly had real)."

Apparently you are successfully converting characters to double 
precision floating-point numbers.

Hope this helps,

Rui Barradas


Em 31-07-2013 21:47, bruce087 at umn.edu escreveu:

            

      
      
      
    

  
  


  


      

    

    

      
      

        


  

        

      David L Carlson
    

    

      
      #
    

  


  

      
It is hard to understand that your R code will not work with a
double matrix since double is just short for double precision
floating point matrix. Your only alternative would be integer.

            

      
      
      
    

        
"It is a historical anomaly that R has two names for its
floating-point vectors, double and numeric (and formerly had
real).

"double is the name of the type. numeric is the name of the
mode and also of the implicit class."

-------------------------------------
David L Carlson
Associate Professor of Anthropology
Texas A&M University
College Station, TX 77840-4352

-----Original Message-----
From: r-help-bounces at r-project.org
[mailto:r-help-bounces at r-project.org] On Behalf Of
bruce087 at umn.edu
Sent: Wednesday, July 31, 2013 3:48 PM
To: r-help at r-project.org
Subject: [R] double matrix?

Hi-

I have a 37 X 473971 character matrix that I am trying to
convert into a 
numeric matrix. When I use the code:

 class(matrix) = "numeric"  

I end up with something called a "double matrix" whose
dimensions are still 
37 X 473971

I have also tried 

new = apply(matrix,2, as.numeric) and got the same thing.

The analysis code I am ultimately attempting to run on this
data requires 
that it be in a numerical matrix, and it is really not okay
with a double 
matrix.

Does anyone know how to fix this?

Thanks.

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      William Dunlap
    

    

      
      #
    

  


  

      
In R "double" and "numeric" mean essentially the same thing.  I think
you are fine. (What called the result a "double matrix"?)

  > z <- cbind(c("11", "12"), c("3.14", "2.718"))
  > str(z)
   chr [1:2, 1:2] "11" "12" "3.14" "2.718"
  > class(z)
  [1] "matrix"
  >
  > class(z) <- "numeric"
  > str(z)
   num [1:2, 1:2] 11 12 3.14 2.72
  > class(z)
  [1] "matrix"
  > z
       [,1]  [,2]
  [1,]   11 3.140
  [2,]   12 2.718
  > log(z)
           [,1]      [,2]
  [1,] 2.397895 1.1442228
  [2,] 2.484907 0.9998963

R "numeric" vectors consist of C "double" or Fortran "double precision"
or "real*8" values - 8 byte double precision floating point numbers with
52 binary digits of precision. 

S supported 4-byte single precision vectors which it also considered
"numeric".

Bill Dunlap
Spotfire, TIBCO Software
wdunlap tibco.com