split special merged text and number in text file

Hi I've special text file (without any spaces or , between text and numbers) (attached file) and I want to split text and number in each lines for example: I have 'X004986Y002446' and I want to convert it to X=004986 and Y=002446 as two variable(Vector/Matrix) how to input text file and split each lines?

thanks

Answer 1

[Stephen23]

Using textscan, fgetl, ftell and fseek in a while loop makes this easy. It will copy the lines without X...Y... verbatim to the new file, and imports the X & Y values into MATLAB for you to process. These are then saved into the new file.

fmt = 'X%dY%d';
fi1 = fopen('G1.txt','rt');
fi2 = fopen('G2.txt','wt');
while ~feof(fi1)
  byt = ftell(fi1);
  txt = fgetl(fi1);
  if strncmp(txt,'X',1)
    fseek(fi1,byt,'bof');
    C = textscan(fi1,fmt);
    X = C{1};
    Y = C{2};
    ... your code here
    M = [X,Y];
    fprintf(fi2,'X%06dY%06d\n',M.');
  else
    fprintf(fi2,'%s\n',txt);
  end
end
fclose(fi1);
fclose(fi2);

Within the if statement you can use txt to identify the current tool block, should you need to do so.

My code was tested on this file (and generates a new file exactly the same!):

[Will Bellknap]

Hi Stephen. your solution is great, but I've some mistake " Excuse Me" another text files may have not X or Y in some lines, it means that X or Y doesn't change from last point. for example:

X00819136Y02363738
X00869936

(means X00869936Y02363738)

I attach another text file

What about this?

[Stephen23]

There might be other ways to resolve this non-uniform file format, but one workaround is to convert it into a uniform file format, for example:

fi1 = fopen('P1.txt','rt');
C = textscan(fi1,'%[^\n]');
fclose(fi1);
C = regexprep(C{1},{'^(X\d+)$','^(Y\d+)$'},{'$1YNaN','XNaN$1'});
fi3 = fopen('P3.txt','wt');
fprintf(fi3,'%s\n',C{:});
fclose(fi3);

Then you can use the same code as my answer give to read the numeric data. How you deal with the NaN' is up to you, e.g. you could change the numeric format string from %d to %f to keep the NaN's as NaN's.

Answer 2

[KSSV]

fid = fopen('data.txt') ;
S = textscan(fid,'%s','delimiter','\n') ;
fclose(fid) ;
S = S{1} ;
S = S(7:end-4) ;
out=regexp(S,'[\d]+','match') ;

out = [out{:}] ;
out = reshape(out',2,[])' ;
out=cell2mat(cellfun(@(x) str2double(x),out,'un',0)) ;

There would be more elegant way..

[Will Bellknap]

Hello. thanks for your good solution. Actually this text file is g-code for a cnc machine and T01 means tool 1 and T02 means tool 2. I have same text files with more tools.

Is it possible to scan such this text files X....Y.... for each tool(T01,T02,...) and after changing X...Y... coordinates (I know how to do it) save it in another text file with as same as input file? ( only X....Y.... changes other part of text not changes) It's a small university project "non-commercial"

[Stephen23]

@Will Bellknap: I have given an answer that shows how to import each block of X & Y values, and creates a new file. It works with any number of T01, T02, ... blocks of data.