how do I extract subsets from a vector

Hello,

I have a vector of the size [70397,2]. It has a title of 2 rows and some text 4 rows between the each data set (2x1801 each). This means a total of 39 sets.

I neet to extract the second column set into a matrix of [39,1801] size.

What would be the most efficient way to do it (I have 100 data sets similar to this one to process)

Thank you

[Dyuman Joshi]

It's not clear what your data is like. It would be helpful if you attach the data or give a sample of what your data looks like.

[Robert Jones]

here are some excerpts from the data file

file starts below. line 3 is an empty line. 1801 is the numberof lines per data set needed to be extracted ==========

# TICRA DATA EXPORT

# 2023-01-17T06:32:48

# Curve: 23.7 GHz → E_lhc → 0.0 deg [Imaginary]

# From : imported_working_1/ff_sphcut_C_feed_sub_main_allpoints_CP.cut

1801

0.00E+00 -6.54E-09

1.00E-01 -2.00E-02

2.00E-01 -7.96E-02

3.00E-01 -1.78E-01

4.00E-01 -3.15E-01

5.00E-01 -4.87E-01

6.00E-01 -6.94E-01

7.00E-01 -9.32E-01

8.00E-01 -1.20E+00

9.00E-01 -1.49E+00

1.00E+00 -1.80E+00

1.10E+00 -2.14E+00

1.20E+00 -2.48E+00

1.30E+00 -2.84E+00

1.40E+00 -3.19E+00

1.50E+00 -3.55E+00

1.60E+00 -3.91E+00

1.70E+00 -4.25E+00

1.80E+00 -4.58E+00

..

...

# Curve: 25.7 GHz → E_lhc → 0.0 deg [Imaginary]

# From : imported_working_1/ff_sphcut_C_feed_sub_main_allpoints_CP.cut

1801

0.00E+00 -6.54E-09

1.00E-01 -2.00E-02

2.00E-01 -7.96E-02

3.00E-01 -1.78E-01

...

...

# Curve: 23.7 GHz → E_lhc → 0.0 deg [Imaginary]

# From : imported_working_1/ff_sphcut_C_feed_sub_main_allpoints_CP.cut

1801

0.00E+00 -6.54E-09

1.00E-01 -2.00E-02

2.00E-01 -7.96E-02

3.00E-01 -1.78E-01

...

...

[dpb]

Attach the file as a file with the paperclip instead of pasting text...you could excerpt only a few sets, whether there are 2 or 2000 sets won't matter for testing.

Answer 1

[dpb]

One way just using the form of the file...

S=readlines('yourfile.txt');                        % import as string array
ixFrom=find(startsWith(S,"# From:"));               % index to beginning each section
sizeSection=str2double(S(ixFrom(1)+1));             % size of each group -- all must be same per Q? spec
isFrom=ixFrom+2;                                    % adjust start of offset from header, count records
for i=1:numel(isFrom)
  i1=isFrom(i); i2=i1+sizeSection-1;                % start, stop records each section
  tmp=str2double(split(S(i1:i2)));                  % convert to numeric array
  if i==1
    A=tmp;                                          % save first set; keep time as well for first set
  else
    A=[A tmp(:,2)];                                 % append second column subsequent
  end
end

Caution, air code, watch out for mismatched/missing paren's, etc., etc., etc., ...

This will return one extra column over the requested; if you really, really don't want the time(?) data as well at all, then just remove the special case and start off with

A=[];

The above uses dynamic reallocation; for no larger files than these the performance hit won't be bad compared to trying to reallocate. Or, you could use a cell array for the intermediary and then cell2mat

ADDENDUM:

For early releases predating readlines, use

S=string(textread('yourfile.txt','%s','delimiter','\n','whitespace',''));

[Robert Jones]

Thanks, will try

[Robert Jones]

I am using version 9.7.0.1190202 (R2019b)

Using your code, I get this error

'Unrecognized function or variable 'readlines'.'

[Dyuman Joshi]

readlines() was introduced in R2020b, won't work for you

[dpb]

Didn't notice the release, sorry...in that case take an intermediary step first...

S=string(textread('yourfile.txt', '%s', 'delimiter', '\n','whitespace', ''));

The editor will complain that textread is not recommended, use textscan instead, but while textscan is somewhat more powerful, it's more of a pain to use because it doesn't accept a filename; you first have to open a file handle with fopen and then fclose it when done.

The above brings in the file a cellstr() array, then converts that to the string array to be consistent with remaining existent code posted.