How do I read in a series of numbers using “Textscan” in MATLAB if the file is mostly text?

Question

I have a text file that have a string of 3 numbers that I need to read into MATLAB.

For Example:

#######################
#
#
#    Text Text Text
#
#
#######################

Blah blah blah = ####
Blah blah blah = ####
Blah blah blah = ####
Blah blah blah = ####
Blah blah blah = ####
Blah blah blah = ####


I_NEED_THIS_STRING =  1234.5 6789.0 1234.5 !Comment blah blah blah

I need to read in those 3 numbers into an array.

PLEASE HELP.

Thanks

Answer 1

If most of the file is irrelevant to your application, I suggest preprocessing with your favorite scripting language or command line tool to find the relevant lines and use textscan() on that.

e.g., from a shell prompt:

grep ^I_NEED_THIS_STRING infile > outfile

in matlab:

fid = fopen('outfile');
C = textscan(fid, 'I_NEED_THIS_STRING = %f %f %f')
fclose(fid)

See the textscan documentation for more details.

Answer 2

An alternative is to use IMPORTDATA to read the entire file into a cell array of strings (with one line per cell), then use STRMATCH to find the cell that contains the string 'I_NEED_THIS_STRING', then use SSCANF to extract the 3 values from that cell:

>> data = importdata('mostly_useless_text.txt','\n');  %# Load the data
>> index = strmatch('I_NEED_THIS_STRING',data);  %# Find the index of the cell
                                                 %#   containing the string
>> values = sscanf(data{index},'I_NEED_THIS_STRING = %f %f %f')  %# Read values

values =

  1.0e+003 *

    1.2345
    6.7890
    1.2345

If the file potentially has a lot of useless text before or after the line you are interested in, then you may use up a lot of memory in MATLAB by loading it all into a variable. You can avoid this by loading and parsing one line at a time using a loop and the function FGETS:

fid = fopen('mostly_useless_text.txt','r');  %# Open the file
newLine = fgets(fid);                        %# Get the first line
while newLine ~= -1                          %# While EOF hasn't been reached
  if strmatch('I_NEED_THIS_STRING',newLine)  %# Test for a match
    values = sscanf(newLine,'I_NEED_THIS_STRING = %f %f %f');  %# Read values
    break                                    %# Exit the loop
  end
  newLine = fgets(fid);                      %# Get the next line
end
fclose(fid);                                 %# Close the file