MATLAB Answers

# Extracting numerical results from a text file at different positions

4 views (last 30 days)
GS76 on 10 Mar 2019
Commented: GS76 on 11 Mar 2019
I have a number of output files to process, written out from Finite Element Analysis software, from which I need to extract data.
I need to extract data under 2 different "headings" from each of the files. However, each of the files has these headings at different rows. I have attached a "zipped" file with 2 files indicating this problem.
The following screenshots show the "headings" highlighted and the numerical data to be extracted, shown inside a "box":
First Heading
Second Heading
I have tried a number of options in Matlab, but I am overwhelmed as I am new to Matlab and programming.
Any assistance and guidance, would be much appreciated.

#### 1 Comment

darova on 10 Mar 2019
read about fileread, regexp

Sign in to comment.

### Accepted Answer

GT on 11 Mar 2019
good catch, my bad:) it was late at night. This should work. In either case if you are learning MATLAB, regexp are very powerful. They exist in many languages, it is worth understanding them:)
a = fileread('0295_PhD_AB~Analysis 1.txt');
% find ( H A R M O N I C = \d )
[b,c] = regexp(a,'( H A R M O N I C = \d )','tokens');
result = table();
% find the 2nd and the 6th number in the table
for i = 1:length(c)
aux = (a(c(i)+586:c(i)+700));
d = regexp(aux,'([0-9.E-+]*)','tokens');
result = [result;table(repmat(b{i},2,1),[d{2};d{6}])];
end
result.Var2 = str2double(result.Var2);

#### 1 Comment

GS76 on 11 Mar 2019
Thank you GT!
This worked out excellently.
I will continue learning and practising "regexp". Thank you for all your help and tips.

Sign in to comment.

### More Answers (1)

GT on 10 Mar 2019
There is probably an easier way to do this... but this should help. (I am using R2018b).
a = fileread('0295_PhD_AB~Analysis 1.txt');
% find ( H A R M O N I C = \d )
[b,c] = regexp(a,'( H A R M O N I C = \d )','tokens');
result = table();
% find
for i = 1:length(c)
aux = (a(c(i)+586:c(i)+700));
d = regexp(aux,'([0-9.E-]*)','tokens');
result = [result;table(repmat(b{i},2,1),[d{2};d{7}])];
end
result

#### 2 Comments

GT on 10 Mar 2019
You can always convert the last column into doubles...
result.Var2 = str2double(result.Var2);
GS76 on 11 Mar 2019
Hi GT,
Thank you very much for this help. This is exactly what I was hoping to achieve.
However, I found the following error whereby the code is extracting the last value from the second column and not the first, when extracting for "HARMONIC = 2". Please see the attached screenshot:
Could you assist me to correct this small problem in the code.
Thank you again for your assistance.

Sign in to comment.

Sign in to answer this question.

R2018b

Translated by