MATLAB: How to count the number of times a string occurs in a cell array in a vectorized format

Question

I have a vector, 'A', of length 'n' with non-unique strings. Some entries will repeat but not necessarily all. I would like to create a cell array that lists the unique strings and another cell array that lists how many times each string occured in vector 'A'.

I would like to perform this in a vectorized method as the cell array is quite large and using a FOR loop to compare each unique entry against the entire vector is very time consuming.

Answer 1

Applying the UNIQUE, STRCMP, and REPMAT functions, finding the number of string repetitions in cell arrays can be performed in a vectorized format.

Using the UNIQUE function, one can obtain cell vector, 'unique_names', of length 'm' from vector 'original_names'. The contents of 'unique_names' will be every string that appears at least once.

unique_names = unique(original_names);

Then using the REPMAT function, one can make a matrix, 'X', where the rows are vector 'original_names' repeated 'm' times resulting in an mxn matrix.

X = repmat(original_names,length(unique_names),1);

Similarly, we make a matrix, Y, with vector 'unique_names' repeated over the columns 'n' times, again, creating an mxn matrix.

Y =repmat(unique_names,1,size(original_names,2))

Next, one can use the STRCMP function to compare matrix 'X' and 'Y'. Taking the sum along the rows will give the quantity of times a given string in the cell array occurred.

quantity = num2cell(sum(strcmp(X,Y),2)');

Please see the attached file for a test sample.

Alternatively you can use the HIST-function to count how many times each string occured in the input vector 'A'.

[a b c] = unique(A)
d = hist(c,length(a))

a contains the the same vector A, but without repetitions.

d contains the count of repetitions.