complicated formatted text with textscan

Dear collegues, I am complicated reading a .txt with a peculiar format. This .txt file is an output of an R function, which I am running with vanilla inside matlab.... So i need to load this txt file programatically.

This is a part of the contents of the text file :

Reservoir_age_offset_14C_yrs_ranges at 90% confidence intervals

Identification_number:  1 
MinRange	MaxRange	probability
-3962 	-1741 	90 	
Mode	MidRange	Median
-2709 	-2852 	-2761 	

Identification_number:  2 
MinRange	MaxRange	probability
-9770 	-8066 	90 	
Mode	MidRange	Median
-8898 	-8918 	-8913 	

Identification_number:  3 
MinRange	MaxRange	probability
-3762 	4226 	87.6 	
4335 	4699 	2.4 	
Mode	MidRange	Median
-559 	468 	305 	

Identification_number:  4 
MinRange	MaxRange	probability
-401 	2224 	90 	
Mode	MidRange	Median
593 	911 	838 	

Identification_number:  5 
MinRange	MaxRange	probability
-709 	976 	90 	
Mode	MidRange	Median
307 	134 	262 	

etc etc etc .................................................

So, to read this text format I implemented this script, the idea is to reconstruct a matrix from the numerical values in the text file

clear
here=pwd;
filename= 'output.txt';%ths is the text file
fileID2 = fopen(filename);
k=1;
Vale=[];
for j=1:50
s=cell2mat(textscan(fileID2,'%f',3,'headerlines',j,'collectoutput',1));

    if isempty(s)==0 %rescue only valid solutions
    Vale(:,k)=s; %brrrrrrppp!!
    k=k+1;
    end

end

fclose(fileID2);

The problem is that I cant read all the numerical values to reconstruct,

it works when I change the number of headerlines manually

e.g.

%s1=cell2mat(textscan(fileID2,'%f',3,'headerlines',7,'collectoutput',1))

However the loop provided incomplete nuber of values

Any idea or more elegant solution with this issue is highly appreciated

Thanks in advance

Jules

[Stephen23]

Is the block of data for Identification_number: 3 really supposed to have two rows of numeric values?:

This makes the block data format irregular, and likely difficult to parse and store.

Otherwise this task can be achieved quite easily using TEXTSCAN (see attached files).

[Jules Ray]

Thank you Stephen, with textscan is quite more easy... I would like to tag this comment as accepted answer but seems you have to post it below ....

thanks again

best

JR

Answer 1

[Stephen23]

I removed the extra line of numbers from block 3, the modified data file is attached.

This code imports the modified data file:

opt = {'CollectOutput',true, 'MultipleDelimsAsOne',true, ...
    'HeaderLines',1,  'EndOfLine',':', 'Whitespace',' \b\t\n\r'};
fmt = '%f%*s%*s%*s%f%f%f%*s%*s%*s%f%f%f%*s';
fid = fopen('data.txt','rt');
tmp = textscan(fid,fmt,opt{:});
fclose(fid);
mat = tmp{1}

mat = 5×7

1 -3962 -1741 90 -2709 -2852 -2761 2 -9770 -8066 90 -8898 -8918 -8913 3 -3762 4226 87.6 -559 468 305 4 -401 2224 90 593 911 838 5 -709 976 90 307 134 262

[Jules Ray]

wonderful, thak you very much