Extracting a number from an input (text file)?

in the middle of my program, I need to read a Number from a file (text or DAT). This is a part of the file

##$SUBNAM9= <"">
##$SW= 2776.7173944802
##$SWIBOX= (0..15)
0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0
##$SW_h= 833333.333333333
##$SWfinal= 0
##$TD= 16384
##$TD0= 1
##$TE= 300
##$TE2= 300

I want Matlab to first read this file, then search for the number in front of "##$TD= ", which is 16384 in this example. Then

TD = 16384;

Thanks in advance

[Amin]

Is it possible at all?

Answer 1

[Walter Roberson]

T = regexp(TheString, '^##\$TD=\s*(.*) \s*$', 'tokens');

Then T{1} will be the string of the number. str2double(T{1}) if you want the numeric value.

[Walter Roberson]

If you have the input as a cell array of strings, then

linematches = strncmp(TheCellString,'%##$TD',5);

and then TheCellString{linematches} would be the full matching line. If you extract positions 7 to end and str2double() that, you would have the number.

[Amin]

Could you please help me again,

when I run string to double, I get this:

>> str2double(T{1})

??? Index exceeds matrix dimensions.

[Walter Roberson]

If the index exceeds the matrix dimensions the implication is that it found no matches.

What form is your TheString ? For regexp() it should either be a cell array of strings, or a single long string containing linefeeds (and possibly carriage returns) to mark the end of individual lines.

The code I showed does not do the reading from file. You could use it as

TheString = fileread('YourFile.txt');

T = regexp(TheString, '^##\$TD=\s*(.*) \s*$', 'tokens');

if isempty(T)

disp('string was not present')

else

disp(str2double(T{1}))

end

[Amin]

In fact I don't know what form my TheString is! :(

the complete code gives me this result:

string was not present

[Oleg Komarov]

If you don't know that you want the number after:

##$TD=

then what are you asking for?

The question is wether the number follows "##$TD= " or "##$TD =" or something else.

[Amin]

I think it is clear in my question that the number follows "##$TD= "

[Walter Roberson]

Darn default behaviours...

T = regexp(TheString, '(?-s)(?m)^##\$TD=\s*(\S+)\s*$', 'tokens');

The exact pattern to use depends on whether you have lines that have embedded spaces and on whether you need to trim the trailing spaces from such lines.

[Amin]

Thanks Walter. Now working. Great!

Answer 2

[Laura Proctor]

I'll put my answer here since it's a little different, but Walter's response is definitely a lot more elegant.

fid = fopen('data.txt');
dataText = fgetl(fid);
while ~feof(fid)
    if strfind(dataText,'##$TD')
        ldata = textscan(dataText,'##$%s %f');
        TD = ldata{2};
        break
    end
    dataText = fgetl(fid);
end
fclose(fid);

[Walter Roberson]

strfind() used in that way would find ##$TD anywhere in the string, but the following textscan() would only match if the line started with $$# . The difference could potentially be a problem -- e.g., if the user had happened to use a comment that included ##$TD as part of it.

strncmp(dataText,'##$TD',5) would be more robust than strfind() in this matter.

[Amin]

This solution is much more understandable for basic people like me. But the 'one line' code from Walter Roberson is so elegant.

I use strncmp(dataText,'##$TD',5) instead of strfind() for this. Thank you Laura Proctor.