Reading a massive file but skipping several lines / rows at a fixed interval

What is the most efficient way to read / analyze rows 1-1990 of an array, then skip 9 lines, then read rows 2000-3990, skip 9 rows, read rows 4000-5990, and so on to the end of the array? Is there a way to vectorize that?

[KSSV]

What information your file has?

[L'O.G.]

It is an array of double precision numeric data

[KSSV]

You can read the entire data and then from these delete the indices you want.

[L'O.G.]

Yes, but how? I know how to delete every 4th row, for example, would be A(4:4:end,:) = [] but I don't know how to apply that for my case where I need to skip / delete multiply lines.

[Stephen23]

Use a loop. Inside the loop you can easily tell MATLAB which part of the file to read, e.g.:

• TEXTSCAN lets you specify how many times the format is applied and how many header lines.
• READTABLE et al let you specify the data location, header location, etc.

What is so special about the rows you wan to ignore: are they intermediate headers or just data you don't want? The answer to that question will change how you can approach this task.

Answer 1

[KSSV]

id = 1:9 ;  % line numbers 
V = 1990:2000:9990 ;  % give the end wisely 
idx = id'+V ;  % make indices 
idx = idx(:) ;  % make indices a column 
% Let A be your array 
A(idx) = [] ;    % remove the lines using the indices 

[L'O.G.]

Thanks. Very nice. One small typo: the last line should be: A(idx,:) = [];

[KSSV]

Yes, if it is an array use A(idx), if it is a matrix, A(idx,:). I have considered it as an array. :)

Answer 2

[Arif Hoq]

try this loop:

A=(1:9990)';　% making an array
idx=2000;   % taken value from 2000
count=(size(A,1)+10)/idx;
first_index=A(1:1990);  % first index value to make the loop simple
C=cell(count,1);
for i=2:count
    C{i}=A(idx*(i-1):(idx*i)-10);    
end
mat=[C{:}]