MATLAB: How to read a text file containing different number of values on each line using MATLAB 7.8 (R2009a)

dataMATLABregexpsplitstringtokens

I am attempting to read data from a text file using MATLAB 7.8 (R2009a). Each line of the text file contains tokens which in this case are numbers separated by one or more spaces. Additionally, each line of the text file contains different number of tokens. As there are a different number of tokens on each line, I am unable to use the TEXTSCAN or FSCANF functions to read the data. Currently, I am reading data from the file element by element.

Is there a faster way to read data from this file?

Best Answer

The code snippet below illustrates how you can read data from a text file containing different number of tokens in every line of the file:

fp = fopen('myDataFile.txt', 'rt');
iter = 1;
while feof(fp) == 0    
      % Read the entire line of data
      strInput = fgetl(fp);
      % Remove any white spaces after the last number on the line. 
      strInput = regexprep(strInput, ' *$', '');
      strNums = regexp(strInput, ' *', 'split');
      numValues = cellfun(@str2num, strNums)
end
fclose(fp);

The code provided above will only work for versions MATLAB 7.5 (R2007b) and higher as the SPLIT option for the function REGEXP was introduced only in MATLAB 7.5 (R2007b).

For earlier versions, replace the following lines of code:

strNums = regexp(strInput, ' *', 'split');
numValues = cellfun(@str2num, strNums)

with

numValues = splitString(strInput)

where the function SPLITSTRING is available for download below.

Related Solutions

MATLAB: Is it possible to insert or replace text in the middle of an ASCII file in MATLAB 8.1 (R2013a)

The ability to replace lines in the middle of an ASCII file is not available in MATLAB 8.1 (R2013a)?

To work around this issue, you can write a new file as you are reading in the original file line-by-line. When you reach the line you would like to change, write the new line to the new file. Then continue to copy the rest of the original file line-by-line to the new file.

An example of this approach that uses regular expressions to identify lines for replacement follows:

% open file for reading
fidr = fopen('TSr.txt','r') ;
% open file for writing
fidw = fopen('TSw.txt','w') ;
% while end of file has not been reached
while ( ~feof(fidr) )
      % read line from reading file
      str = fgets(fidr) ;
      % match line to regular expression to determine if replacement needed
      match = regexp(str,'(?<=Line )\d(?= is this.)', 'match' ) ;
      % if line is to be replaced
      if ( ~isempty(match) )
          % define replacement line
          str = ['This is line ',match{1},'.',char(10)] ;
      end
      % write line to writing file
      fwrite(fidw,str) ;
end

where the 'TSr.txt' file that is read is

This is line 1.
This is line 2.
Line 3 is this.
Line 4 is this.
This is line 5.

and the 'TSw.txt' file that is written is

This is line 1.
This is line 2.
This is line 3.
This is line 4.
This is line 5.

MATLAB: How to scan text file

filename = 'AppropriateName.txt';
S = fileread(filename);
L = regexp(S, '\r?\n', 'split');
data = cellfun(@(s) str2double(regexp(s, ',', 'split')), L, 'uniform', 0);

The result will be a cell array with one entry per line, with each entry being the numeric data from the corresponding line.

Best Answer

Related Solutions

MATLAB: Is it possible to insert or replace text in the middle of an ASCII file in MATLAB 8.1 (R2013a)

MATLAB: How to scan text file

Related Question