Extracting field values from cell array of structures

I have a cell array of structures. Each structure has the fields : Start, End, Id..

RP_Cell{1,1}.Start = [1 2 3; 4 6 4] 
RP_Cell{1,1}.End= [11 7 33; 4 16 9] 
RP_Cell{1,1}.Id= [23;-4] 

RP_Cell{1,2}.Start = [1 -2 4] 
RP_Cell{1,2}.End= [17 6 3] 
RP_Cell{1,2}.Id= [5] 
.
.
RP_Cell{2,1}.Start = [1 2 87; 4 23 1; 7 3 1] 
RP_Cell{2,1}.End= [15 65 21; 35 1 6; 6 3 1] 
RP_Cell{2,1}.Id= [9; 34; 56] 
.
.
.
.

Now, I want to have the data in the fields separately - Start_List, End_List

Start_List{1,1} = [1 2 3; 4 6 4] 
Start_List{1,2} = [1 -2 4]
.
.
Start_List{2,1} = [1 2 87; 4 23 1; 7 3 1]
. 

End_List{1,1} = [11 7 33; 4 16 9] 
End_List{1,2} = [17 6 3] 
.
.
End_List{2,1} = [15 65 21; 35 1 6; 6 3 1] 
.

How do I do this in the simplest way without using loops

[Adam]

Why do you so specifically not want to use loops?

[Amulya NV]

I just wanted to know if there is a simple one shot way of doing such operations, because I was reading some posts where a oneliner will do operations like these.

Also, the problem is that I am getting this cell array from inside of a parfor loop. I am new to parfor and cell arrays so I thought its a little tricky for me with the 2D indices when inside a parfor.

Answer 1

[Guillaume]

Assuming that all cells of the RP_Cell cell array only contain a scalar structure with the same fields in the same order, you can transform the cell array into a structure array:

RP_Array = cell2mat(RP_Cell);

It is then trivial to transform that structure array into a cell array with struct2cell.

fnames = fieldnames(RP_Array);
Allfields = struct2cell(RP_Array);  %Creates a numfield x M x N cell array
%Allfields(1, :, :) corresponds to RP_Array(:, :).(fnames{1})
%Allfields(2, :, :) corresponds to RP_Array(:, :).(fnames{2})
%etc.
Start_List = squeeze(Allfields(strcmp('Start', fieldnames(RP_Array)), :, :));
End_List = squeeze(Allfields(strcmp('End', fieldnames(RP_Array)), :, :));

Note that if you know for sure that Start is the first field of the structure, you can dispense with the strcmp and simply use:

Start_List = squeeze(Allfields(1, :, :));

Using strcmp is safer, since if the order ever happen to change, it will find the correct field.

[Amulya NV]

I am unable to do the cell2mat() operation as my each cell has the same structure with same fields but with different number of entries

Something like the following:

1*1 struct | 1*1 struct | 1*2 struct |
1*2 struct | 1*1 struct | 1*1 struct |
1*4 struct | 1*1 struct | 1*3 struct |

[Guillaume]

So if, RP_cell{1,3} is a 1x2 struct array, e.g.

RP_Cell{1, 3}(1) = struct('Start', [1 2 3;4 5 6], 'End', [10 12; 15 17], 'Id', 4);
RP_Cell{1, 3}(2) = struct('Start', [-1 -2 -3], 'End', [21 23], 'Id', [12;15]);

What should Start_List{1, 3} consists of?

Possibly, you want

Start_List = cellfun(@(c) {c.Start}, RP_Cell, 'UniformOutput', false);

but this will create a cell arrays of cell arrays since you need to store more than one matrix per cell of Start_List.

[Amulya NV]

Thank you Guillaume.. This is what I wanted :)

Apologies for the confusing question