Splitting up large amount of data into smaller tables based on one coloumn

clear all; close all;

% x = x, y = y, r= raman shift, i = intensity

[x] = textread('rawData.txt');

y=size(x);

rv=length(x);% rv = raman shift value

B1=zeros(rv,4);

for c=1:1:rv-1;

for i=1:4;

if x(c,3)>x(c+1,3)

B1(c,i)=x(c,i);

B1=B1+1;

elseif x(c,3)<x(c+1,3);

B1(c,i)=x(c,i);

B1=B1+1;

end

end

end

Hi all this is my current code, i have a very large data set comprised of 4 coloumns, and i need to seperate it when a value in the third coloumn is smaller than the next value. The number of values between this happening is not consistent, but i need to seperate it out into lots of tables of the 4 coloumns

Any help would be greatly appreciated!!

Answer 1

[Bob Thompson]

Hmm, I think I kind of understand what you are trying to do. Let me know if this works.

mark = 1;
cll = 1;
for i = 2:size(x,1)
    if x(i,3) > x(i-1,3);
        tests{cll} = x(mark:i,:);
        cll = cll + 1;
        mark = i;
    end
end

This should look for the different breaks, and record the different data sets into cells in 'tests'.

[pillowheaded]

Thank you so much that has done exactly what i need it to!

[pillowheaded]

Hi

I have been using the code you wrote and realised that it misses out the last block of data is there anyway to add additional lines of code to ensure it copies all the data?

thank you for your help

Answer 2

[Peter Perkins]

You have not said why you want to do that, but in general, you DON'T want to split one table up into lots of tables. You will usually find that less convenient than just haivng an indicator variable in your one table.

In any case, to create that indicator, you don't need loops. Find the changepoints, and cumsum:

>> t = array2table(rand(10,3))
t =
  10×3 table
      Var1        Var2        Var3  
    _________    _______    ________
      0.10665    0.43141     0.85303
       0.9619    0.91065     0.62206
    0.0046342    0.18185     0.35095
      0.77491     0.2638     0.51325
       0.8173    0.14554     0.40181
      0.86869    0.13607    0.075967
     0.084436    0.86929     0.23992
      0.39978     0.5797     0.12332
      0.25987    0.54986     0.18391
      0.80007    0.14495     0.23995
>> t.Indicator = cumsum([true; t.Var3(2:end) > t.Var3(1:end-1)])
t =
  10×4 table
      Var1        Var2        Var3      Indicator
    _________    _______    ________    _________
      0.10665    0.43141     0.85303        1    
       0.9619    0.91065     0.62206        1    
    0.0046342    0.18185     0.35095        1    
      0.77491     0.2638     0.51325        2    
       0.8173    0.14554     0.40181        2    
      0.86869    0.13607    0.075967        2    
     0.084436    0.86929     0.23992        3    
      0.39978     0.5797     0.12332        3    
      0.25987    0.54986     0.18391        4    
      0.80007    0.14495     0.23995        5    

[pillowheaded]

Hi thank you for your answer,

the reason i want to do this is my data is a series of scans taken at various points but instead of the data being individual files the output is they are all stacked into one giant table with 4 coloumns.

i see what you have done and i like it but it is more benifical to me to have them split up as i then go on to fit peaks as each set of data is a spectre.

how would i use this code to then divide up the data?

[Peter Perkins]

n = 5;
c = cell(n,1);
for i = 1:n
    c{i} = t(t.Indicator==i,1:3);
end