How to read a .txt file?

Is it possible to read this file line by line and create two matrices with two columns and a different number of lines?

The file looks like this and it's created with a function for storing some fingerprint data and associating it with a name and birth date:

And I want to create a matrix with the values of X and Y on two columns for both types.

Let me translate those lines:

Nume - Name; 
Data - Birth Date; 
Numar de sfarsituri de crestatura - Number of ridge endings; 
Numar de bifurcatii - Number of bifurcations; 
Sfarsituri de crestatura - Ridge Endings; 
Bifurcatii - Bifurcations;

As I said, every ridge ending or bifurcation has two coordinates (they're in a matrix, of course). For example, lets say I want the ridge endings thingy. I need to read those lines and print a matrix like this: SFC = [84 32;78 60;131 105; ... ]

Is that possible?

[Stephen23]

@Graur Alin: It makes our lives much easier if you give us text data as text, not as an image. Then we can actually use your text, instead of having to write it all out again.

You can either include it as code (formatted using the {} Code button), or upload it using the paperclip button (then pressing both Choose file and Attach file).

Answer 1

[Azzi Abdelmalek]

fid=fopen('fic.txt')
l=fgetl(fid)
k=1
while ischar(l)
  r{k}=l
  l=fgetl(fid)
  k=k+1;
end
fclose(fid);
ii=find(~cellfun(@isempty,regexpi(r,'x\sy')));
idx=ii+1;
idy=[ii(2:end)-1 numel(r)];
M=regexpi(r(idx(1):idy(1)),'\d+','match');
jj=~cellfun(@isempty,M);
M=M(jj);
M=reshape(str2double([M{:}]),2,[])'
N=regexpi(r(idx(2):idy(2)),'\d+','match');
jj=~cellfun(@isempty,N);
N=N(jj);
N=reshape(str2double([N{:}]),2,[])'

[Graur Alin]

r looks ok. The conversion it's troublesome:

Columns 1 through 4

      [1x68 char]    'Nume: Graur_Alin '    'Data: 01_06_1992 '    [1x38 char]

    Columns 5 through 9

      [1x24 char]    [1x68 char]    [1x68 char]    [1x27 char]    [1x68 char]

    Columns 10 through 14

      ' X         Y  '    ' 84     31'    ' 78     60'    '131    105'    '112    130'

    Columns 15 through 19

      '142    150'    '105    154'    ' 80    160'    '130    168'    [1x68 char]

    Columns 20 through 24

      'Bifurcatii : '    [1x68 char]    ' X         Y  '    ' 87     88'    ' 72     92'

    Columns 25 through 29

      ' 64     94'    ' 45     96'    '105    102'    ' 61    140'    ' 92    141'

    Columns 30 through 32

      '104    142'    '130    143'    '109    147'

It works, somehow.

I get this error though:

f_read
Index exceeds matrix dimensions.

Error in f_read (line 13)
M = regexpi(r(idx(1):idy(1)),'\d+','match');

[Azzi Abdelmalek]

attach the text file you want to read

[Graur Alin]

The thing is I have no experience with text files in MatLab at all.

[Azzi Abdelmalek]

fid=fopen('fic.txt')
l=fgetl(fid)
k=1
while ischar(l)
  r{k}=l
  l=fgetl(fid)
  k=k+1;
end
fclose(fid);
ii=find(~cellfun(@isempty,regexpi(r,'\sx\s+y')))
idx=ii+1;
idy=[ii(2:end)-1 numel(r)];
M=regexpi(r(idx(1):idy(1)),'\d+','match');
jj=~cellfun(@isempty,M);
M=M(jj);
M=reshape(str2double([M{:}]),2,[])'
N=regexpi(r(idx(2):idy(2)),'\d+','match');
jj=~cellfun(@isempty,N);
N=N(jj);
N=reshape(str2double([N{:}]),2,[])'

[Graur Alin]

Thank you, it works!

[Graur Alin]

One more question if you don't mind. I've got that function's input parameters as a file path and it returns me M and N and now I have to read multiple text files in a loop and compare them to my fingerprint image.

I use this first:

files = dir('*.txt');
for m=1:length(files)
   [M,N] = pc_read (??);

What do I put there ("??") if my read file function's input looks like this:

function [M N] = pc_read(file)
fid = fopen(file);

Thank you!

Answer 2

[Stephen23]

For faster and neater code you can use textscan, which also converts the strings to numeric values where appropriate. You can try something like this:

fid = fopen('test.txt','rt'); % specify the permission!
hdr = textscan(fid,'%s%s',4,'Delimiter',':','HeaderLines',1);
tx1 = textscan(fid,'%s',1,'Whitespace','-:\n');
xy1 = textscan(fid,'%f%f','MultipleDelimsAsOne',true,'HeaderLines',1);
tx2 = textscan(fid,'%s',1,'Whitespace','-:\n');
xy2 = textscan(fid,'%f%f','MultipleDelimsAsOne',true,'HeaderLines',1);
fclose(fid);
%
hdr = [hdr{:}];
tx1 = [tx1{:}];
tx2 = [tx2{:}];
xy1 = cell2mat(xy1);
xy2 = cell2mat(xy2);

And we get the following values:

>> hdr
hdr = 
    'Nume'                                 'Graur_Alin'
    'Data'                                 '01_06_1992'
    'Numar de sfarsituri de crestatura'    '8'         
    'Numar de bifurcatii'                  '10'        
>> tx1
tx1 = 
    'Sfarsituri de crestatura '
>> xy1
xy1 =
    84    31
    78    60
   131   105
   112   130
   142   150
   105   154
    80   160
   130   168
>> tx2
tx2 = 
    'Bifurcatii '
>> xy2
xy2 =
    87    88
    72    92
    64    94
    45    96
   105   102
    61   140
    92   140
   104   142

To perform this in a loop you can use this:

files = dir('*.txt');
names = {files.name};
for k = =1:length(names)
    file = names{k}
    ...code here
end

Or read more here:

http://de.mathworks.com/matlabcentral/answers/57446-faq-how-can-i-process-a-sequence-of-files

Because there was no sample file uploaded with the original questrion I created my own, which you can find here: