How to separate characters from words and write both of them to matrice when reading from text file

example I have: ",considering);" I want this like

A{1}{1} = , 
A{1}{2} = considering 
A{1}{3} = ) 
A{1}{4} = ;

note: there is no white space between characters and words

[per isakson]

I assume that you look for a code that will handle more strings than this example. What is the rule according to which the input string shall be split?

[Tural]

actually I have to split text so non-letter characters will be stored in its cell ,when I do textscan (considering) is word considering stored with brackets

Answer 1

[Geoff Hayes]

Tural - you could try using regexp to split the string into its letter and non-letter parts, splitting on the non-letters. Something like

 str = ',considering);';
 [A, nonLetterIdcs] = regexp(str,'[^A-Za-z]','split');

In the above, we split the str message on those characters that are not (the ^ means "not") in the square bracket expression (so we split on anything that is not in the range A-Z or a-z), and our output becomes

 A = 
    ''    'considering'    ''    ''

 nonLetterIdcs =
     1    13    14

A is a cell array, and nonLetterIdcs is an array of indices of the non-letters that we have split on. Now we need to combine the two as

 A(cellfun(@(x)isempty(x),A)) = cellstr(str(nonLetterIdcs)');

We use cellfun to look for the empty strings of A which correspond to the positions in the original str that have non-letters. And we need to "fill" these empty strings with the corresponding non-letter from str, using the nonLetterIdcs. Having done that, A becomes

 A = 
    ','    'considering'    ')'    ';'

Answer 2

[per isakson]

A construct based on regexp:

    >> A{1} = regexp( ',considering);', '\W|\w++', 'match' );

    >> A{1}{2}
    ans =
    considering