reading a text file to matlab with different formatting

Question

Hi I have a text file, but unfortunately its been poorly formatted, however i want to read the content of the text file into a a matrix, but I dont know how to do that. the when try to use fscanf, textscan,textread and the rest it just copies everything into one cell, but i dont want it that way. this how the content looks like: so i want to read only the decimals not the absolute figures. cans someone help me.

1 : 13.27 ; 3 : 20.68 ; 6 : 8.271 ; 7 : 3.308 ; 8 : 8.328 ; 9 : 6.655 ; 10 : 16.58 ; 11 : 9.925 ; 12 : 12.41 ; 13 : 4.135 ; 14 : 9.925 ; 15 : 11.58 ; 16 : 10.87 ; 17 : 1.654 ; 18 : 4.962 ; 19 : 6.655 ; 22 : 10.98 ; 23 : 24.25 ; 24 : 47.33 ; 25 : 11.6 ; 26 : 9.925 ; 27 : 5.809 ; 28 : 5.001 ; 29 : 6.617 ; 30 : 7.577 ; 31 : 9.155 ; 32 : 7.444 ; 33 : 28.58 ; 34 : 9.155 ; 35 : 35.83 ;

Answer 1

conceptually:

take the string.

Change : and ; with spaces and \n (return) with ;

Assign the string to a variable, the variable should become a matrix.

With a dual loop change the value of each elemnt with

var(i,j)=floor(var(i,j)) - var(i,j);

Here you are (If I didn't misunderstood you).

Answer 2

I assume that the colon (:) seperates values in a row and the semicolon (;) is separating rows. With this assumption the following function should read your data into a MATLAB matrix

function dat = readData(filename)
% FUNCTION dat = readData(filename)
% Reads data from a nonstandard formatted file, filename
% INPUTS:
%    filename:  Full path to data file with the following format
%                      1 : 2; 3 : 4
%               where the : separates values in a row and ; separates rows

% open the file for reading
fid = fopen(filename);

ii=0; % row index

while ~feof(fid) % loop until we find the end of the file
    str = fgetl(fid); % get a line of text

    while 1
        % split the string into its component parts
        % assume that values are split with colon ":" and
        % rows are identified with a
        % semicolon ";".
        % 
        % split at the first row using strtok
        [rowStr rem]=strtok(str,';');

        % split rowStr using colon
        [str1,str2]=strtok(rowStr,':');

        % get rid of the colon in str2
        str2 = strrep(str2,':',' ');

        str1 =strtrim(str1);
        str2 =strtrim(str2);

        % store data if we found any
        if ~isempty(str1)&& ~ isempty(str2)
            ii=ii+1; % increment row index
            dat(ii,:) = [str2double(str1) str2double(str2)];
        end

        if isempty(rem), break; end
        str = rem(2:end);
    end
end
fclose(fid);

You can than use the functions round, floor, ceil to extract the 'decimal values'.

Answer 3

Just using textscan and ignoring the things you don't need, like the numbers and : gives you quite a simple solution:

fid = fopen('test.txt', 'rt');
data = textscan(fid, '%*u %*1s %f', 'Delimiter', ';');
fclose(fid);

Change test.txt to your filename. data is a cell with your doubles in it.

>> data{:}

ans =

   13.2700
   20.6800
    8.2710
    3.3080
    8.3280
    6.6550
   16.5800
    9.9250
   12.4100
    4.1350
    9.9250
   11.5800
   10.8700
    1.6540
    4.9620
    6.6550
   10.9800
   24.2500
   47.3300
   11.6000
    9.9250
    5.8090
    5.0010
    6.6170
    7.5770
    9.1550
    7.4440
   28.5800
    9.1550
   35.8300