Re: request for help with textread

[Kristoffer Walker]

Philip,

First off, thank you for the quick response, and on the weekend. 
Secondly, thanks so much for developing textread.  It is an extremely 
useful program.

I'm not brave, so I'll work with the stock strread.  But I will endeavor 
to install 3.6.4 for the workaround you pointed out.  However, when I 
first tried to install octave on Linux Mint, the default installation 
was version 3.2 if I recall correctly.  This version did not have 
textread.  So after googling around, I learned to do this to install 3.6:

sudo apt-add-repository -y ppa:picaso/octave
sudo apt-get update
sudo apt-get install octave
sudo apt-get install liboctave-dev

That installed version 3.6.1.  Is there a similar set of commands that 
will enable me to install 3.6.4?  If not, maybe compiling it from source 
is not a painful experience?

Thank you,
Kris


On 09/07/2013 07:09 AM, PhilipNienhuis wrote:
> Kristoffer Walker wrote
> > Hello,
> >
> > I have recently installed octave 3.6.1 on Linux Mint v. 13 (version of
> > Ubuntu 12).  Data file "test.in":
> >
> > 1336 901
> > 0.000000 0.000000 0.3329580775 -0.0022473728 -0.0008216525
> > 0.000000 0.200000 0.3320154915 -0.0023726505 -0.0009491121
> >
> > [x10,y10]=textread("test.in","%f %f%*[\n]","headerlines",1);
> > error: textread: A(I): index out of bounds; value 1 out of bound 0
> > error: called from:
> > error:   /usr/share/octave/3.6.1/m/io/textread.m at line 75, column 3
> > <snip>
> > Can someone tell me what I am doing wrong?  I read the help page
> > completely.
>
> Maybe first of all you installed a too old version of Octave?
> The stable version has been 3.6.4 for quite a while. I think the bug (index
> out-of-bound) has been fixed since 3.6.1, I vagely remember that.
>
> But anyway even with 3.6.4 and the development version it doesn't work.
> After having put in some spaces for readability, textread.m shows me the
> actual problem:
>
> > > [x10, y10] = textread ("test.in", "%f %f%*[\n]", "headerlines", 1)
> error: strread: %q, %c, %[] or bit width format specifiers are not supported
>   yet.
> error: called from:
> error:   X:\Octave\octave-3.7.6+_MXE\share\octave\3.7.6+\m\io\strread.m at l
> ine 308, column 7
> error:   X:\Octave\octave-3.7.6+_MXE\share\octave\3.7.6+\m\io\textread.m at
> line 180, column 31
> > >
>
> ...which means that the %*[\n] format conversion specifier isn't supported.
> Are you trying code written for Matlab? (seems so)
>
> If you skip the rightmost 3 columns by using the "%*f" format, textread
> works for the datafile you showed:
>
> > > [x10, y10] = textread ("test.in", "%f %f %*f %*f %*f", "headerlines", 1)
> x10 =
>
>     0
>     0
>
> y10 =
>
>     0.00000
>     0.20000
>
> > >
>
> I you're brave, you could also try a patched strread.m (the work horse
> behind textread.m) where I had implemented some of those missing format
> conversion specifiers (notably [] and *[]) but with a lot of gotchas; I
> don't remember clearly, it was quite a while ago. See here:
>
> https://savannah.gnu.org/patch/?7817
>
> I don't remember what versions of textread.m the patched strread.m will work
> with - might be from around early 2012. Later versions probably catch the
> "unsupported" format specifiers before handing the job to strread.m
> You'll have to try yourself if you go this route.
>
> Philip
>
>
>
>
> --
> View this message in context: 
> http://octave.1599824.n4.nabble.com/request-for-help-with-textread-tp4657102p4657110.html
> Sent from the Octave - General mailing list archive at Nabble.com.
> _______________________________________________
> Help-octave mailing list
> address@hidden
> https://mailman.cae.wisc.edu/listinfo/help-octave