loading text file to matrix without delimiters

I have a text file dataset with Y number of lines and each line has exactly 250 characters (including spaces).

I want to put each character (even if it is a space) into a matrix so I create a matrix with 250 columns and Y rows. We can replace the spaces with NaN.

I have been playing with textscan and few other functions but cannot seem to get it. Anyone have any ideas?

Answer 1

[Jos (10584)]

So each digit is a single value, and spaces are to be replaced with NaN …

M = char(textread('example.txt','%s','delimiter','')) - '0' ;
M(M==(' '-'0')) = NaN ; % replaces spaces with NaN's
% M is now a 4-b-250 numerical array

btw the second line of example.txt is 251 characters long

Answer 2

[Azzi Abdelmalek]

d=importdata('file.txt')

[Ammar]

Importdata creates the correct number of rows but it uses spaces as a delimiter to create only a few columns.

So for: 4334___34343_455656435356345__4534

where _ = space

it would create 4 columns. Meanwhile, I want each character or space to be put into a separate column

[Azzi Abdelmalek]

This is not clear. Do you want to import numeric data or what? Post the four line of your file

[Ammar]

I have attached an example file. There are 4 lines Each line has 250 characters (including spaces)

I want to turn this into a matrix with 250 columns and 4 rows

Answer 3

[Image Analyst]

When I saved your example it didn't have exactly 250 characters on every line. I had 251 in the second line for some reason and the last line had only 1 character. So I made the code a little more robust than you might need it, but extra robustness never hurts.

fid = fopen('example.txt');
tline = fgetl(fid);
lineCounter = 1;
charArray = tline;
fprintf('%d characters in line #%d: %s\n', length(tline), lineCounter, tline)
while ischar(tline)
  tline = fgetl(fid);
  if length(tline) < 3
    continue;
  end
  fprintf('%d characters in line #%d: %s\n', length(tline), lineCounter, tline)
  charArray = [charArray; tline(1:250)];
  lineCounter = lineCounter + 1;
end
fclose(fid);
% Display in command window.
charArray

Of course you can get rid of the display lines (fprintf, etc.) if you want.

Answer 4

[Ammar]

Thank you everybody!! I had been banging my head on the wall for a few hours on this item.

Answer 5

[Joseph Cheng]

simple method

fid = fopen('example.txt','r');
line = fgetl(fid);
spaces = strfind(line,' ');
line(spaces)=0;
x=line(:);
y=hex2dec(x)';
y(spaces)=NaN;

this is just an example but given the line of text use strfind(line,' ') to find the spaces. Note these indexes to be replaced by NaNs, make the spaces 0 and transpose it to a 250x1 array such that hex2dec will convert each into a number, then replace the index of the found spaces to NaN.

I used hex2dec as i noticed in your example you have a 'C' in there. So unless it was a typo its a hex number? if it is a typo then just use str2num or str2double instead

Answer 6

[Ammar]

Instead of putting one character per cell is there a way to specifcy which cell they go into?

For example, if I have text like the following

24359_435934009____________90909

where _=space

And I want to put it into a matrix such that

characters 1-5 go into column 1 (a.k.a 24359)

character 6 goes into column 2 (ak.a. NaN)

characters 7-10 goes into column 3

and so on....

[Azzi Abdelmalek]

Is this an answer or a question?

[Ammar]

new question (Variation to the question)

[Joseph Cheng]

use my method i had and modify it.

fid = fopen('example.txt','r');
line = fgetl(fid);
spaces = strfind(line,' ');
spaces = [0 spaces length(line)+1];
cellpos = 1;
for i=1:length(spaces)-1
   portion = line(spaces(i):spaces(i+1)-1);
   if portion ==' '
    blah{cellpos}=NaN;
else
blah{cellps}=portion;
end

[Azzi Abdelmalek]

Ammar, this is considered as an answer, you can add comments by clicking on [comment on this answer]* or create a new question