# Sort Cell Array after accumarray function

5 views (last 30 days)
MakM on 26 Dec 2022
Commented: MakM on 27 Dec 2022
I have following code;
row_f=({0;0;0;1;2;3;4;0});
row_s=({'a';'a';'b';'b';'c';'c';'a';'b'});
t={'12/09/2022 04:28:01 PM';'12/09/2022 04:28:02 PM';'12/09/2022 03:28:03 PM';'12/09/2022 03:28:04 PM';'12/09/2022 02:28:05 PM';'12/09/2022 02:28:06 PM';'12/09/2022 04:28:03 PM';'12/09/2022 03:28:05 PM'};
all=[row_f, row_s,t];
[~,~,X] = unique(all(:,2));
dataset= accumarray(X,1:size(all,1),[],@(r){all(r,:)})
dataset = 3×1 cell array
{3×3 cell} {3×3 cell} {2×3 cell}
disp(dataset{1,1});
{[0]} {'a'} {'12/09/2022 04:28:01 PM'} {[0]} {'a'} {'12/09/2022 04:28:02 PM'} {[4]} {'a'} {'12/09/2022 04:28:03 PM'}
disp(dataset{2,1});
{[0]} {'b'} {'12/09/2022 03:28:03 PM'} {[1]} {'b'} {'12/09/2022 03:28:04 PM'} {[0]} {'b'} {'12/09/2022 03:28:05 PM'}
disp(dataset{3,1});
{[2]} {'c'} {'12/09/2022 02:28:05 PM'} {[3]} {'c'} {'12/09/2022 02:28:06 PM'}
If we see the timestamp in the third column, timestamp in the first cell (dataset{1,1}) is greater than dataset{2,1} and dataset{3,1}; How do I sort the timestamp acc to ascending order after accumarray function so I can have the new dataset cell values as dataset{3,1} in the first cell because its timestamp is smaller. then dataset{2,1} in the second row cell and finally the datatset{1,1} in the thirst row cell.
like the new value for the dataset would be:
dataset = 3x1 cell array
{2x3 cell}
{3x3 cell}
{3x3 cell}
Matt J on 26 Dec 2022
Edited: Matt J on 26 Dec 2022
If, as in your example, the time stamp is always the same for a given alphabetic label in column 2, it is not clear why you need both.
If they are not always the same, it is not clear how you want them sorted in that case.

Matt J on 26 Dec 2022
Edited: Matt J on 26 Dec 2022
row_f=[0;0;0;1;2;3;4;0];
row_s=({'a';'a';'b';'b';'c';'c';'a';'b'});
t={'12/09/2022 04:28:01 PM';'12/09/2022 04:28:02 PM';'12/09/2022 03:28:03 PM';'12/09/2022 03:28:04 PM';'12/09/2022 02:28:05 PM';'12/09/2022 02:28:06 PM';'12/09/2022 04:28:03 PM';'12/09/2022 03:28:05 PM'};
T=sortrows( table(row_f(:), row_s(:),t(:)) , [3,2]);
[~,~,id]=unique(T{:,2},'stable');
dataset=splitapply(@(x){x},table2cell(T),id); dataset{:}
ans = 2×3 cell array
{[2]} {'c'} {'12/09/2022 02:28:05 PM'} {[3]} {'c'} {'12/09/2022 02:28:06 PM'}
ans = 3×3 cell array
{[0]} {'b'} {'12/09/2022 03:28:03 PM'} {[1]} {'b'} {'12/09/2022 03:28:04 PM'} {[0]} {'b'} {'12/09/2022 03:28:05 PM'}
ans = 3×3 cell array
{[0]} {'a'} {'12/09/2022 04:28:01 PM'} {[0]} {'a'} {'12/09/2022 04:28:02 PM'} {[4]} {'a'} {'12/09/2022 04:28:03 PM'}

the cyclist on 26 Dec 2022
Edited: the cyclist on 26 Dec 2022
Here is one way:
row_f=({0;0;0;1;2;3;4;0});
row_s=({'a';'a';'b';'b';'c';'c';'a';'b'});
t={'12/09/2022 04:28:01 PM';'12/09/2022 04:28:02 PM';'12/09/2022 03:28:03 PM';'12/09/2022 03:28:04 PM';'12/09/2022 02:28:05 PM';'12/09/2022 02:28:06 PM';'12/09/2022 04:28:03 PM';'12/09/2022 03:28:05 PM'};
all=[row_f, row_s,t];
[~,~,X] = unique(all(:,2));
dataset= accumarray(X,1:size(all,1),[],@(r){all(r,:)})
dataset = 3×1 cell array
{3×3 cell} {3×3 cell} {2×3 cell}
% Sort it
[~,sortIndex] = sort(cellfun(@(x)x(1,end),dataset));
datasetSorted = dataset(sortIndex,:)
datasetSorted = 3×1 cell array
{2×3 cell} {3×3 cell} {3×3 cell}
MakM on 27 Dec 2022
Thanks cyclist for the answer, it worked too