Re: fscanf transposes matrix

[Paul Kienzle]

Octave follows Matlab's broken implementation of scanf, which fills
the matrix by column rather than by row.  E.g.,

> sscanf("0 1 2\n3 4 5","%f",[2 3])
ans =

  0  2  4
  1  3  5

This is not what you would expect from looking at the file:

  0  1  2
  3  4  5

When using fscanf to read from a file, you will invariably do so
like

> fscanf(fid,"%f",[nc,nr])'

so it is consistent to call it like this, but with nr == Inf.

The case when you have a fixed number of rows but an unknown number of
columns you must handle with something like:

> x = fscanf(fid,"%f"); x = reshape(x,length(x)/nr,nr)'

Unfortunately the behaviour of scanf can't be fixed without breaking
compatibility with existing Octave code and with ported Matlab code, both
of which depends on this behaviour.

Paul Kienzle
address@hidden

On Wed, Jan 30, 2002 at 03:39:43PM +0100, Martijn Brouwer wrote:
> Hi,
> I am (quite) new to octave and I am looking for a way to read an entire
> matrix of measurement data from a file at once.
> I now use a=fscanf(file,"%f%f",[2,Inf]) to read a file containing two
> columns of numbers, but the resulting matrix is transposed compared to
> the file. Offcourse, this is easily cured with a ', but it is counter
> intuitive. Reading with [Inf,2] would in my opinion be better, but this
> results in an one column matrix.
> 
> Any comments?
> 
> Bye,
> 
> Martijn Brouwer
> 
> 
> 
> 
> 
> -------------------------------------------------------------
> Octave is freely available under the terms of the GNU GPL.
> 
> Octave's home on the web:  http://www.octave.org
> How to fund new projects:  http://www.octave.org/funding.html
> Subscription information:  http://www.octave.org/archive.html
> -------------------------------------------------------------
> 



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------