How to automatically import data out of blocks in ascii files?

I have some data from measurements and the data files are like:

WL 1200.00
x1 tab y1
x2 tab y2
.
.
.
WL 1201.00
x1 tab y1
.
.
.
Wl 1202.00
.
.

Every block starts with 'WL' and then a wavelength. Then beneath the x-y data devided by tabs.

The problem is that the data arrays do not always have the same length. How can I tell Matlab that a new block starts with 'WL'. My goal is to get a 3D diagram with all the curves for all the different wavelengths...

Answer 1

[Cedric]

Several solutions would be available. Here is one which uses some of the most basic tools for reading ASCII file content:

 fid = fopen('myData.txt', 'r') ;
 content = {} ;
 cId     = 0 ;                                     % Content/block counter/ID.
 while ~feof(fid)
    line = fgetl(fid) ;
    if isempty(line),  continue ;  end
    if line(1) == 'W'                              % Create new block; save 
        cId = cId + 1 ;                            %  WL, init data array.
        content{cId}.WL = sscanf(line, '%*s %f') ;
        content{cId}.data = [] ;
    else                                           % Append xy to data array.
        xy = sscanf(line, '%f\t%f') ;
        content{cId}.data = [content{cId}.data; xy.'] ;
    end
 end
 fclose(fid) ;

With that and the following data

 WL 1200.00
 1.8 2.3
 3   8
 WL 1200.01
 2.6 9.0
 3.4 1.2
 8   7
 7.3 5.4

you get

 >> length(content)
 ans =
     2

 >> content{1}
 ans = 
      WL: 1200
    data: [2x2 double]

 >> content{1}.data
 ans =
    1.8000    2.3000
    3.0000    8.0000

 >> content{2}.data
 ans =
    2.6000    9.0000
    3.4000    1.2000
    8.0000    7.0000
    7.3000    5.4000

Here is a mixed REGEXP/SSCANF solution, just for the fun of it (could also be STRFIND/SSCANF). It is likely to be faster than the first version on large files..

 buffer  = fileread('myData.txt') ;
 blocks  = regexp(buffer, 'WL', 'split') ;
 blocks  = blocks(2:end) ;                        % Skip 1st empty block.
 nBlocks = length(blocks) ;
 content_regexp = cell(nBlocks, 1) ;
 for k = 1 : nBlocks
    values = sscanf(blocks{k}, '%f', Inf) ;
    content_regexp{k}.WL = values(1) ;
    content_regexp{k}.data = reshape(values(2:end), 2, []).' ;
 end