From octave-sources-request at bevo dot che dot wisc dot edu  Mon Nov  8 08:07:57 1999
Subject: Reading ascii data file with comments
From: Vladimir Eltsov <ve at boojum dot hut dot fi>
To: octave-sources at bevo dot che dot wisc dot edu
Date: Mon, 08 Nov 1999 16:07:54 +0200

This is a multi-part message in MIME format.
--------------5C8B22527A624BDEB6962C51
Content-Type: text/plain; charset=koi8-r
Content-Transfer-Encoding: 7bit

Hello,

Here is octave m-file which reads data from ascii file. Unlike built-in
'load' function it allows to ignore arbitrary comments or to collect
them in a string output argument. Also it can distribute data columns to
different arrays for you. 

This program uses no octave functionality beyond 2.0.x, but
unfortunately it doesn't work there due to octave bugs. So version later
than 2.1.20 is required.

Regards,

   Vladimir Eltsov
--------------5C8B22527A624BDEB6962C51
Content-Type: text/plain; charset=koi8-r;
 name="read_ascii_data.m"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="read_ascii_data.m"

function [...] = read_ascii_data (file, incols, comchar)

# usage: x = read_ascii_data (file, ncols)
#        [x, comstr] = read_ascii_data (file, ncols, comchar)
#        [x_1, x_2, ..., x_ncols] = read_ascii_data (file)
#        [x_1, x_2, ..., x_ncols, comstr] = read_ascii_data (file, comchar)
#
# This function read the table with 'ncols' columns from
# ascii file. Either string for a file name or file id is passed
# through 'file' parameter. If file name is "-" then 
# stdin is used. Result is placed to matrix 'x'
# or each column is placed to x_i. Non-parseable input is 
# ignored. Those lines which starts with 'comchar' char
# are collected in 'comstr' output argument.

### Note: this function is usable only with octave versions > 2.1.20
  
# Vladimir Eltsov
# veltsov at mail dot ru
# 8 November 1999     

switch (nargin)
  case (1)
    ncols = nargout;
    cm = 0;
    cc = 0;
  case (2)
    if (isstr(incols))
      ncols = nargout-1;
      cm = 0;
      cs = incols(1);
      cc = 1;
    else
      ncols = incols;
      cm = 1;
      cc = 0;
    endif
  case (3)
    ncols = incols;
    cm = 1;
    cs = comchar(1);
    cc = 1;
  otherwise
    usage("read_ascii_data(file [,ncols] [,comment_char])");
endswitch

ncols = max(ncols,1); # if result goes to ans, then nargout==0
     
do_close = 0;
if (isstr(file))
  if (strcmp(file,"-"))
    fin = stdin;
  else
    fin = fopen(file,"r");
    if (fin<0)
      error(["read_ascii_data: cannot open input file ", file]);
    endif
    do_close = 1;
  endif
else
  if (! is_scalar(file) && ! is_stream(file))
    error("read_ascii_data: 'file' parameter should be string or file id");
  endif
  fin = file;
endif

y = [];
cb = "";

save_eleo = empty_list_elements_ok;
unwind_protect
  empty_list_elements_ok = 1;
  while (! feof(fin))
    [z, k] = fscanf(fin,"%f",Inf);
    if (k>0)
      y = [y; z];
    endif
    s = fgets(fin);
    if (isstr(s) && cc && s(1)==cs)
        cb = [cb s];
    endif
  endwhile
unwind_protect_cleanup
  empty_list_elements_ok = save_eleo;
end_unwind_protect

if (do_close)
  fclose(fin);
endif

k = length(y);
c = rem(k,ncols);
if (c != 0)
  warning("read_ascii_data: non-exact number of rows were read");
  y = [y; zeros(ncols-c,1)];
endif
x = reshape(y,ncols,length(y)/ncols)';

if (cm)
  vr_val(x);
else
  for i=1:ncols
    vr_val(x(:,i))
  endfor
endif
if (cc)
  vr_val(cb);
endif

endfunction

--------------5C8B22527A624BDEB6962C51--