How do I remove duplicates from a structure array?

Suppose I create the following structure:

    animals(1).name='Northern Mockingbird';
    animals(1).class='Bird';
    animals(2).name='Cheetah';
    animals(2).class='mammal';
    animals(3).name='Wolf';
    animals(3).class='mammal';
    animals(4).name='Eagle';
    animals(4).class='Bird';
    animals(5).name='Turtle';
    animals(5).class='Reptile';
    animals(6).name='Bee';
    animals(6).class='Insect';
    animals(7).name='Chameleon';
    animals(7).class='Reptile';
    animals(8).name='Barn Spider';
    animals(8).class='Arachnid';
    animals(9).name='Ant';
    animals(9).class='Insect';
    animals(10).name='Northern Cardinal';
    animals(10).class='Bird';

Then I set animals(11) equal to animals(4), then I sort the whole thing through a littany of conversions/reshapings, the sortrows command and another littany of conversions/reshapings.

Is there some way to remove the duplicate entries from that array of structures?

Answer 1

[Azzi Abdelmalek]

a= {animals.name}
b= {animals.class}
c=cellfun(@(x,y) [x y],a', b','un',0)
[ii,ii]=unique(c,'stable')
animals=animals(ii)

[Azzi Abdelmalek]

%or

[ii,ii]=unique({animals.name},'stable')
animals=animals(ii)

Because if the name is the same, the class should be the same.

[Craig]

Agreed. The actual structure array is file names and their respective paths; I used the animal array because it is easier to check, the difference (that I had not thought of) is that there could be files that have identical names but reside in different folders. Since your first solution looks for unique name/class pairs (through concatenation), it will still work for file names and paths.

'c=cellfun(@(x,y) [x y],a', b','un',0)' was baffling to me. I sort of understand the 'cellfun' and am slowly getting acquainted with the '@' symbol, but the '(x,y) [x y]' part was confusing to me at first. I think I get it now though. I couldn't find the 'un' property in the documentation; is it the same as 'UniformOutput'?

Answer 2

[Jonathan Sullivan]

It's not pretty, but this should work:

isUnique = true(size(animals));

for ii = 1:length(animals)-1
    for jj = ii+1:length(animals)
        if isequal(animals(ii),animals(jj))
            isUnique(ii) = false;
            break;
        end
    end
end

animals(~isUnique) = [];

[Craig]

Thanks for the quick answer. I was hoping to avoid for loops, though I will admit that the loops you came up with are way prettier than I had thought a for loop solution would look like.

[John King]

Wow, only had to schange the name from "animals" to my struct name and this worked for me! Thank you Jonathan!