How can i read different data types (strings numbers vectors) from one .txt-File?

At the moment i am using a .txt - File for my data input. (Only numbers with 3 columns)

In this .txt-file I want also store all the relevant formating data for my diagram. -> see attachment file. So i can use this parameters in my matlab code. How can i read them?

This was my old matlab code just with my Data Input:

Input_Matrix = textread('InputXX.txt')
[zeilen,spalten]=size(Input_Matrix)
x = Input_Matrix(:,1)
y = Input_Matrix(:,2)
z = Input_Matrix(:,3)

Thank you guys. Sincelery Philipp

[Guillaume]

Since the format of the file is one of your own creation you have no choice but to write your own parser.

You could either read the whole file at once and parse it using regular expressions or parse the file as you read it using a combination of low level I/O functions and textscan.

[Philipp Mueller]

sounds very complicated :/ I dont understand.. Is there a predifend format?

[Guillaume]

There are some existing formats such as xml or json that are text (thus human readable / writable) and that matlab can easily parse if you want to be able to write your input data with a text editor. Or you could write a script that generates matlab variables.

Otherwise, the best way to store this sort of things is in a mat file, but of course it's not something that you can write / edit in a text editor.

Note that parsing your file is not that complicated (depending how fault tolerant you want your parser to be).

[Philipp Mueller]

do you have some links for me? thx

Answer 1

[Guillaume]

As per my comment you could use an xml file and xmlread or a json file and jsondecode (the latter being a lot less verbose) but it means a substantial change of syntax for your input file.

It's not that complicated to write a parser for your current format. This would work (although lacking a lot of syntax validation):

diagramoptions = [];
diagramlimits = [];
inputdata = [];

wholecontent = fileread('InputXX.txt');
sections = regexp(wholecontent, '\**([^*]*)\**([^*]*)', 'tokens');
for section = sections
   switch(strtrim(section{1}{1}))
      case 'Diagram Options'
          keyvalues = regexp(section{1}{2}, '([^\n\r=]*)=([^\n\r=]*)', 'tokens');
          diagramoptions = cell2table(vertcat(keyvalues{:}), 'VariableNames', {'Key', 'Value'});
      case 'Diagram Limits'
          header = strsplit(regexp(section{1}{2}, '[^\n\r]*', 'match', 'once'));
          content = textscan(section{1}{2}, repmat('%f', 1, numel(header)), 'HeaderLines', 2);
          diagramlimits = table(content{:}, 'VariableNames', header);
      case 'Input Data'
          inputdata = cell2mat(textscan(section{1}{2}, '%f%f%f', 'HeaderLines', 1));
      otherwise
          warning('Unknown section: %s', section{1}{1});
   end
end

%see the outputs
openvar diagramoptions
openvar diagramlimits 
openvar inputdata

[Philipp Mueller]

Thank you for your great help i really appreciate it. Your code works fine but i want to understand it completely. So i have some questions. I have already used the matlab help function which helped me a little bit to understand your comands. But sometimes i have problems to understand. Here are my questions: (Sorry)

sections = regexp(wholecontent, '\**([^*]*)\**([^*]*)', 'tokens')

I dont understand this part

'\**([^*]*)\**([^*]*)', 'tokens'

Of course i know what does this command do.

Second question:

switch(strtrim(section{1}{1}))

I dont know what "1" means in this command.

Third question:

 keyvalues = regexp(section{1}{2}, '([^\n\r=]*)=([^\n\r=]*)', 'tokens')%\n -> \new line; \r carriage return

What does this many

([^\n\r=]*)=([^\n\r=]* say?

Fourth question:

header = strsplit(regexp(section{1}{2}, '[^\n\r]*', 'match', 'once'))

Thank you.

[Guillaume]

I'm afraid you're going to have to learn the regular expression language to fully understand the code I wrote. Note that there are plenty of tutorials on the web. Regular expressions are extremely useful for parsing text.

The '\**([^*]*)\**([^*]*)' regular expression is designed to match each section of your file starting with an **** xxx ** header and all the following lines up to the next ***. For each match, it extracts two tokens (the xxx and the content of the section). The exact meaning of that regular expression is

• match 0 or more * (the \**). I just realised that this should actually have been \*+, match 1 or more *.
• followed by a token that consist of 0 or more character that is anything but a * (the ([^*]*)). Once again, it should have been ([^*]+), capture one or more character that is not a *
• followed by 0 or more * (the \**), again it would be better if it was \*+
• followed by another capture of 0 or more character which are anything but * (the second ([^*]*)). Again |([^*]+) would be better.

The second regular expression is used to split the key value pairs into two tokens on each side of the =. [^\n\r=]* matches any number of characters that are not line ending or =. Again, a + instead of the * would be better.

The last regular expression is there simply to remove the end of line marker that's left from the first regexp.

sections is a 1 x numberofsections cell array of 1x2 cell arrays (containing the section header and content). The for loop creates a section variable which is a scalar cell array containing just one of these 1x2 cell array. The first {1} is to get that 1x2 cell array, the second {1} is to get the header of that section. Maybe if the code was written:

for sectionindex = 1:numel(sections)
    switch(strtrim(sections{sectionindex}{1}))

it would be clearer for you

[Philipp Mueller]

thanks, i will check it. I just spend a couple of hours with matlab. So thatsis the reason for my stupid questions.

[Philipp Mueller]

@Guillaume ok, i understood thank you :) Last question: For example i have 3 diagrams to plot. That means my inputXX.txt has 9 sections -> 3 times diagram options 3 times input data and 3 times diagram limits. how can i seperate it? Is it possible to use

• openvar diagramoptions[i]
• openvar diagramlimits[i]
• openvar inputdata[i]

i...index -> 1,2 or 3 diagram. Thank you in advance

[Guillaume]

The simplest way would be to append to diagramoptions, diagramlimits, inputdata every time you encounter a new one. It's easily done by replacing the assignment with

diagramoptions{end+1} = cell2table(...  %same for the other two

You can then index each section using cell array indexing:

openvar diagramoptions{1}