HOW TO SORT DATA IN CELL ARRAY

A {1x4}

Such that

A = {A{1,1},A{1,2},A{1,3},A{1,4}}

A{1,1}={}
A{1,2}={0.5	0.2
    1	 2
    3	 4
    NaN	NaN
    5	6
    7	8
    9	10
    NaN	NaN}
A{1,3}={0.5    0.8
    5	6
    7	8
    9	10
    NaN	NaN
    11	12
    13	14
    15	16
    17	18
    NaN	NaN}
A {1,4} = {0.7	0.9
    19	20
    NaN	NaN
    25	26
    27	28
    29	30
    31	32
    NaN	NaN}

I want to separate data when there is NaN NaN in the matrix, want to store values before first NaN NaN in one cell array and after NaN NaN in another cell array.

B1 = {A{1, 2}(1:3,:))}
B2 = {A{1, 3}(1:4,:))}
B3 = {A{1, 4}(1:2,:))}
C1 = {A{1, 2}(5:7,:)}
C2 = {A{1, 3}(6:9,:)}
C3 = {A{1, 4}(4:7,:)}

B = {{B1},{B2},{B}}
C = {{C1},{C2},{C3}}

[Luna]

why are you using cell arrays inside cell arrays? What makes you understand which part belongs to B and which belongs to C?

Answer 1

[Ameer Hamza]

Are you looking for something like this code? Note that I have made some modifications to the datatypes and removed A{1,1} = {}, since it makes the code simpler. This code will work as long as there are two nan rows in the matrix.

A{1,1}=[0.5	0.2
    1	 2
    3	 4
    NaN	NaN
    5	6
    7	8
    9	10
    NaN	NaN];
A{1,2}=[0.5    0.8
    5	6
    7	8
    9	10
    NaN	NaN
    11	12
    13	14
    15	16
    17	18
    NaN	NaN];
A{1,3} = [0.7	0.9
    19	20
    NaN	NaN
    25	26
    27	28
    29	30
    31	32
    NaN	NaN];

nan_rows = cellfun(@(x) find(isnan(x), 2), A, 'UniformOutput', 0);

B = cellfun(@(x, y) x(1:y(1)-1, :), A, nan_rows, 'UniformOutput', 0);
C = cellfun(@(x, y) x(y(1)+1:y(2)-1, :), A, nan_rows, 'UniformOutput', 0);

[Akshay Sahu]

It's not working.

[Ameer Hamza]

Is the code not running?

[Akshay Sahu]

Yes it's showing that isnan is undefined

[Ameer Hamza]

Which MATLAB release are you using. isnan() have been defined since R2006a, and it is a part of MATLAB language itself, you don't need some special toolbox. Maybe there is some issue with MATLAB installation. Try

which isnan

What is the output?

[Akshay Sahu]

[Akshay Sahu]

I'm using Matlab2014

[Ameer Hamza]

I am not sure of the problem, find function is also defined since R2006a. Try

which find

[Guillaume]

looks like a very corrupt matlab installation. Both find and isnan have been built-in functions since forever.

[Akshay Sahu]

I think problem is cell array inside cell array??

because both funtions are working properly on matrix

so now any one has the solution that how can I convert cell array inside a cell array to matrix form

[Ameer Hamza]

To avoid nested cell arrays, define A like I have given in my answer. Or

A = {[0.5 0.2; ....], [0.5 0.8; ...], [...]}

[Akshay Sahu]

As you were saying that, this code will work as long as there are two nan rows in the matrix but what if we don't know number of nan rows then is it possible to save them in different cell array using loops?

[Ameer Hamza]

Try this. B is a cell array of cell arrays. See the result with indexing like this: B{1}{1}, B{1}{2}, B{2}{1}, ....

A{1,1}=[0.5	0.2
    1	 2
    3	 4
    NaN	NaN
    5	6
    7	8
    9	10
    NaN	NaN
    0.5	0.2
    1	 2
    3	 4
    NaN	NaN];
A{1,2}=[0.5    0.8
    5	6
    7	8
    9	10
    NaN	NaN
    11	12
    13	14
    15	16
    17	18
    NaN	NaN
    0.5    0.8
    5	6
    7	8
    9	10
    NaN	NaN];
A{1,3} = [0.7	0.9
    19	20
    NaN	NaN
    25	26
    27	28
    29	30
    31	32
    NaN	NaN
    0.7	0.9
    19	20
    NaN	NaN];

total_nan_rows = sum(isnan(A{1}(:,1)));
nan_rows = cellfun(@(x) find(isnan(x), total_nan_rows), A, 'UniformOutput', 0);

for i=1:total_nan_rows
    if i==1
        B{i} = cellfun(@(x, y) x(1:y(i)-1, :), A, nan_rows, 'UniformOutput', 0);
    elseif i==total_nan_rows
        B{i} = cellfun(@(x, y) x(y(i-1)+1:end-1, :), A, nan_rows, 'UniformOutput', 0);
    else
        B{i} = cellfun(@(x, y) x(y(i-1)+1:y(i)-1, :), A, nan_rows, 'UniformOutput', 0);
    end
end

[Akshay Sahu]

Yeah thanks man Ameer Hamza it's working.