Skip to content

import data from Matlab & error msg when install package "e1071"

3 messages · Yan Yu, Marc Schwartz, David Meyer

Yan Yu
# 
Hello,
  I am trying to import data from Matlab..
  when i looked up R documentation, it says, package "e1071" have command
(read.octave) to import data from octave.
but when I tried to install package by using:
install.packages("e1071");
I got the following message: ( BTW, my platform is linux version 2.4.18-3
my gcc is 2.96).

* Installing *source* package 'e1071' ...
checking for C++ compiler default output... a.out
checking whether the C++ compiler works... yes
checking whether we are cross compiling... no
checking for suffix of executables...
checking for suffix of object files... o
checking whether we are using the GNU C++ compiler... yes
checking whether g++ accepts -g... yes
configure: WARNING: g++ 2.96 cannot reliably be used with this package.
configure: error: Please use a different C++ compiler.
ERROR: configuration failed for package 'e1071'


I am wondering did anyone successfully install "e1071" on linux box, what
gcc are you using?

Or Is there another way to import data from Matlab?

Thanks a lot,
yan
Marc Schwartz
# 
package.
There are known problems with gcc 2.96, which was released with some
distros (ie. Red Hat), but was not an official release.  Long story,
but you can probably Google on "gcc 2.96" and get the history. One
place to start is here: http://gcc.gnu.org/gcc-2.96.html.

You can (should) upgrade to a more recent version, either via your
distro's support network or you can go to the gcc web site at
http://gcc.gnu.org/, to download the latest version, which is 3.2.2.

HTH,

Marc Schwartz
David Meyer
#
Yan Yu wrote:
Well, not `2.96', apparently, which actually does not exist (see former
messages in the mail archives on this topic). This `version' of gcc is
buggy and at least breaks the `svm' code in e1071, that's why this check
has been added. So you are well advised to upgrade your compiler suite
anyway.
Depends on your data. If you only have tabular data and your MATLAB file
uses ascii representation, `read.table' could be your friend. If you
insist on using `read.octave' from e1071 without updating your
C-compiler, just source the `read.octave.R' file from the source
package, it's plain R code and so is not affected by any C compiler. For
more complex data, you might consider using `StatDataML'---the R package
is on CRAN, the corresponding MATLAB functions will appear at

http://www.omegahat.org/StatDataML/

by the end of this week.

Best,
David.