I am trying to read data from a text file (attachment). This is a typical summary output from a CFD simulation which has a short description in the first couple of lines, then has columns of numerical data after that. I have always handled CFD ouput using excel but it seems like working with .TXT files as output is easier and the program that I use for CFD computations generates output in .TXT fromat. However, one issue that I constantly face when dealing with .TXT files is that it is not clear to me in which format the data is read into the machine, how the existing commands hanlde the data, or even which command will be the most suitable one for the data format at hand. The good thing is that the data files will always look like this, so if I knew which command to use, I wouldn't need to switich to another one.
What I need to know is this:
- Which command to use, and I have tried all of these: readtable, readmatrix, textread & textscan but each one has a slight quirk that makes it unsuitable for the kind of search that I need to conduct or it is necessary to use more than one command to deal with both the text and the numerical data. In case of textread, e.g., the data seems to be read in the form of a single column cell array, and NONE of the common string search commands seem to work here (contains, strcmp, strfind).
If I use textread:
InText=textread(InputFile,'%q')
the input data is in the form of a 670x1 cell array and none of the search commands find the string I am looking for.
If, instead, I use textscan, as suggested by MATLAB, the ouput is not shown, but MATLAB indicates that it is a 6X1 cell array.
The search commands again will not work.
- How to find the location of a certain srting in the first lines of the text. In case of the file in attachmentm that string would e.g. be 'Volume Average', but the list may include a handful of more options. This will then determine how the code will proceed with further calculations. I tried
- Location of the headers of the numerical data, as in line 5 of the attached file: 'Eta0 (Pa*s)', 'Velocity magnitude (m/s)', 'abs(u2) (m/s)' and so on. This will then tell the program to treat each numerical column accordingly. It is tempting to just read the numerical part and assume the position of each column, but one should not forget that each user may have saved the output file in a different sequence of columns and therefore the best practice is to fully automate the search for values by finding the assigned name for each column. The good news is that the names are fixed and never change their format.
here is the whole script:
clc; format long g; clear all;
[FN,FP] = uigetfile('*.txt','MultiSelect','on');
disp('No files selcted');
disp(['Files selected:', fullfile(FP,FN)]);
if class(FN) =='char', FN={FN}; end
Text2Find='Volume Average';
InText=textread(InFile,'%q')
ParaIdx=find(contains(InText,Text2Find))
find(strcmp(InText,'Velocity Magnitude' ))
Thanks,
Saeid
