Suppose that we have this cell array of data :
a={43 432 2006;
254 12 2008;
65 35 2000;
64 34 2000;
23 23 2006;
64 2 2010;
32 5 2006;
22 2 2010}
Last column of this cell array is years. I want classify data(rows) based on years like this :
a_2006 = {43 432 2006;
32 5 2006;
32 5 2006}
a_2008 = {254 12 2008};
a_2000 = {65 35 2000;
64 34 2000}
a_2010 = {64 2 2010;
22 2 2010}
I have different years in column three in every cell array (this cell array is a sample) so I want an automatic method to determine the years and classify them to a_yearA
, a_yearB
etc. or other naming that I can distinguish years and call data with years easily in following codes.
How can I do this?
Thanks.da
eval is not recommanded here. I would suggest to save it in a struct array like that
a={43 432 2006;
254 12 2008;
65 35 2000;
64 34 2000;
23 23 2006;
64 2 2010;
32 5 2006;
22 2 2010};
tmp = cell2mat(a);
% clear a or take another variable name for your struct to prevent warnings
clear a
years=unique(tmp(:,3));
for n = 1:numel(years)
a.(['a_' num2str(years(n))])=tmp(tmp(:,3)==years(n),:);
end
looks like that
>> a
a =
a_2000: [2x3 double]
a_2006: [3x3 double]
a_2008: [254 12 2008]
a_2010: [2x3 double]
>> a.a_2000
ans =
65 35 2000
64 34 2000
Furthermore, you can easily access your struct arrays in a loop again (what you can't do such easy with the eval hack).
>> names=fieldnames(a)
names =
'a_2000'
'a_2006'
'a_2008'
'a_2010'
>> for n = 1:numel(names)
a.(names{n})
end
Well, say there are strings in your cell
a={'43' '432' 2006;
'254' '12' 2008;
'65' '35' 2000;
'64' '34' 2000;
'23' '23' 2006;
'64' '2' 2010;
'32' '5' 2006;
'22' '2' 2010};
% save original variable 'a' for accessing later
tmp.cell = a;
% save just the years in a matrix for more easy accessing + we need the
% indizes for getting the values from the cell later in the loop
tmp.years = cell2mat(a(:,3));
% clear variable a or take another variable name for your struct to prevent warnings
clear a
% get unique years for iteration
tmp.unique_years=unique(tmp.years);
for n = 1:numel(tmp.unique_years)
a.(['a_' num2str(tmp.unique_years(n))])=tmp.cell(tmp.years==tmp.unique_years(n),:);
end
The result is the same, but your structs are now cells
>> a
a =
a_2000: {2x3 cell}
a_2006: {3x3 cell}
a_2008: {'254' '12' [2008]}
a_2010: {2x3 cell}