Finding unique set from large dataset

Hello!

How would one go about to find a unique set of variables that can depict shortly which variables are used in a large dataset using a matlab code?

For example if the english dictionary is my large dataset, I want the output to be the 26 alphabets-saying these are the unique variables used in your large dataset.

Another example

If x = {"abc", "bcd", "ded"}

I want the output as {"a","b","c","d","e"}

Thanks in advance!

Answer 1

[madhan ravi]

a = cellfun(@(z)regexp(z,'.','match'),x,'un',0);
Output = num2cell(unique([a{:}]))

[KALYAN ACHARJYA]

+1

@Madhan Is this also we can yes?

x ={"abc","bcd","ded"};
s=convertStringsToChars(strcat(x{:}))
data=num2cell(s);
%Next remove all duplicate cell elemnts 

[madhan ravi]

Thanks Kalyan :)

Perhaps:

z = regexp(strcat(x{:}),'.','match');
Output = num2cell(unique(z)) 

edit:

Kalyan your method is in fact better , didn't verify it properly before.

[Guillaume]

Yes, Kalyan, your approach is actually better. Using a regular expression to extract individual characters is a bit overkill.

A few notes:

The OP shows a cell array of scalar string arrays. That's extremely innefficient and pointless. It should be either a string array:

x =["abc","bcd","dedf"];  %a string array

or a cell array of char vectors:

x = {'abc', 'bcd', 'dedf'};  %cell array of char vector (cellstr)

The string array will use less memory.

If starting with a string array, it can be simply:

x =["abc","bcd","dedf"];  %a string array
unique(char(join(x, '')));

If starting with a cell array of char vector:

x = {'abc', 'bcd', 'dedf'};
unique([x{:}])

[madhan ravi]

Thank you Guillaume :)

[Sanjana Sankar]

Thank you all. I was looking for the output from Guillaume's method. Thanks a lot!!

Answer 2

[Waqar Ali Memon]

Hello Sanjana Sankar,

Refer: https://de.mathworks.com/matlabcentral/answers/470189-counting-the-same-occurance-of-a-row-string

Hope it helps :-)

[Sanjana Sankar]

Thank you!!

Answer 3

[Bruno Luong]

x = ["abc", "bcd", "ded"] % no need using curly bracket for strings

string(unique(cat(2,x{:}))')'

[Sanjana Sankar]

Thanks!