Read Textfile (lines with different formats) line by line

Hello,

I have a text file (lets call it an input file) of this type: My kind of input file % Comment 1 % Comment 2

4 %Parameter F

2.745 5.222 4.888 1.234 %Parameter X

273.15 373.15 1 %Temperature Initial/Final/Step

3.5 %Parameter Y

%Matrix A

1.1 1.3 1 1.05

2.0 1.5 3.1 2.1

1.3 1.2 1.5 1.6

1.3 2.2 1.7 1.4

I need to read this file and save the values as variables or even better as part of different arrays. For example by reading I should obtain Array1.F=4; then Array1.X should be a vector of 3 real numbers, Array2.Y=3.5 then Array2.A is a matrix FxF. There are tons of functions to read from text file but I don't know how to read these kind of different formats. I've used in the past fgetl/fgets to read lines but it reads as strings, I've used fscanf but it reads the whole text file as if it is formatted all equally. However I need something to read sequentially with predefined formats.. I can easily do this with fortran reading line by line because read has a format statement. What is the equivalent in MATLAB?

Best Regards and many thanks in advance Emanuel

Answer 1

[Benjamin Kraus]

There are many different strategies, but I think you are going to have to use regexp at some point.

Within a for loop, you can read one line at a time using fgetl. Once you have the line, it looks like you need to determine where the % is in the line to know whether the data is before the % or if the data is on the next line. You could use find to find the index of the first %:

ind = find(str == '%',1);
if ind == 1
  % Read next line for data
else
  % Data is earlier in the line.
end

To determine the parameter name, I think regexp is the only option. For example:

out = regexp(str,'^(?<value>[^%]*)%(?<type>Parameter|Matrix) (?<name>[\S]+)','once','names')

The output from that statement should be a structure with three fields: value (the part before the %), type (the part immediately after the %), and name (the characters after either Parameter or Matrix). It won't work on the line that includes "Temperature Initial/Final/Step". You can either use a different pattern match for that (for example, you could just use contains or strfind to determine if the word 'Temperature' is in the line), or you can try to modify the expression I proposed. It isn't clear which would be easier.

Once you've separated those pieces, you can work with them all individually. You can use sscanf or str2num to convert the strings into doubles.

The last part you may need is how to use a variable to access a struct field. For example:

fieldname = out.name; % This is the 'F' or 'Y' or 'A' from above.
A.(fieldname) = str2num(out.value); % Store the value in the field of the struct.

That might be enough to get you started.

[Emanuel Crespo]

I don't need to read the name of the parameters they don't change so I can do it directly on the program. So If I understood the way to do it is to read each line with fgetl and then convert to number using the str2num? But the probleme here is in one line I have for example 3 values how do I use a str2num in a line like (2 3 4 %Comment) to save a vector [2 3 4] and ignore the comment?