I have a string array, for instance:
arr = ['hello'; 'world'; 'hello'; 'again'; 'I----'; 'said-'; 'hello'; 'again']
How can I extract the most frequent string, which is 'hello' in this example?
Add a comment
|
2 Answers
First step, use a cell array rather than string array:
arr = {'hello', 'world'; 'hello', 'again'; 'I----', 'said-'; 'hello', 'again'};
Second, use unique to get the unique strings (this doesn't work on a string array, which is why I suggest the cell):
[unique_strings, ~, string_map]=unique(arr);
Then use mode on the string_map variable to find the most common values:
most_common_string=unique_strings(mode(string_map));
6 Comments
Eitan T
+1: but there's no need for cell arrays. You can use
unique(arr, 'rows').
Hugh Nolan
Oh great, thanks! I don't use them very often, didn't know this function.
Hugh Nolan
Just a note about string arrays and the above comment: in this instance, the string would need to be reformatted so that each string was a separate row, rather than trying to have two strings on a single line - this only works as a cell, otherwise Matlab considers the whole line a single concatenated string, i.e. the initial
arr in the question is equivalent to ['helloworld','helloagain';,'I----said-';'helloagain']Eitan T
All strings in the question are concatenated vertically with a semicolon. I think you copy-pasted
arr wrong.Hugh Nolan
Oh weird. Thanks. Don't know how that happened.
|
It is better to use cell arrays and regexp function; the behavior of string arrays may not be what you expect.
arr = {'hello', 'world'; 'hello', 'again'; 'I----', 'said-'; 'hello', 'again'};
If you use
hellos = sum(~cellfun('isempty', regexp(arr, 'hello')));
it will return the number of 'hello''s in cell array arr.
