Asked by Bojan
on 19 Dec 2012

First I using command xlsread to import some data from excel and got vector row (or column it's never mind), then I try to create new matrix using for loops and if command, but I have problem, because I don't know how to do that... :)

Example: I import this

x=[1 5 3 7 10 10 10 15 4 8 3 5 11 19 10 10 10]

I want to create something like this:

a=[1 5 3 7] and b=[15 4 8 3 5 11 19]

,and new x=[mean(a) mean(b)]

It's necessary to use loops because every excel file that I import have different number of cell, for example: data from new excel is

x1=[1 2 3 10 10 10 4 5 6 7 8 9 11 12 13 10 10 10]

That the reason why I need block of code. Number 10 that is repeated is area that separate group of number, and I wanna to use that condition to create new vectors.

Thanks for your help.

*No products are associated with this question.*

Answer by Bojan
on 24 Dec 2012

Accepted answer

Thank everyone for trying to help me, but I find solution by my self. Maybe is little complicated, but it works, and that is most important... :)

I use "xlsread" command to import data from excel and I get matrix x1, then I use for loop to get my new matrix x2.

Code:

x2=zeros(13,7);

% caunters

m=1; n=1; t=1;

for k=m:length(x1) if x1(k,:)~=100 x2(t,n)=x1(k,1); t=t+1; else if (x1(k,:)==100 && (x1(k+1,:)==100 && x1(k+2,:)==100)) m=k+3; n=n+1; t=1; continue end end end

Answer by Image Analyst
on 19 Dec 2012

This is **trivial** if you have the Image Processing Toolbox - it's one line:

segments = regionprops(x1~=10, x1, 'PixelList', 'MeanIntensity'); % Get all your a's, b's, etc.

The above line will get you all the segments (all their values into individual arrays), and the mean values of all the arrays.

Question - is it **always 10** that is the "dividing" zone between segments? Or is it *any* number that repeats?

Show 1 older comment

Image Analyst
on 19 Dec 2012

Bojan
on 21 Dec 2012

Image Analyst
on 21 Dec 2012

By all over the place I meant that the 10 or 100 or whatever can appear in stretches of any length in any position. It's not like you can just assume that each stretch of 100 occurs every 6th element throughout the array. If it did occur with such predictable periodicity you can just use normal, good old indexing to extract out the subarrays. However since your 100's occur at elements 6, 15, 27, etc. in random locations, that's what requires you to use regionprops which can handle that kind of data. The code is the same as I first gave you:

separationNumber = 100; % or 10 or whatever you want. segments = regionprops(x1 ~= separationNumber , x1, 'PixelList', 'MeanIntensity');

Opportunities for recent engineering grads.

## 3 Comments

## Benjamin (view profile)

Direct link to this comment:http://se.mathworks.com/matlabcentral/answers/56999#comment_118129

Can there be more than a and b? In other words, could you need to split x into n arrays?

## Bojan (view profile)

Direct link to this comment:http://se.mathworks.com/matlabcentral/answers/56999#comment_118134

Yes it can. My need function to split x into n arrays. Where number 10 is zone between segments.

## Bojan (view profile)

Direct link to this comment:http://se.mathworks.com/matlabcentral/answers/56999#comment_118140

Yes, I use number 10 for example. This is the imported vector from .xls file:

x1=[mean(a) mean(b)]