views:

832

answers:

1

Using regexp with tokens on cell array of strings I've got cell array of cells. Here is simplified example:

S = {'string 1';'string 2';'string 3'};
res = regexp(S,'(\d)','tokens')
res = 

    {1x1 cell}
    {1x1 cell}
    {1x1 cell}
res{2}{1}
ans = 
    '2'

I know I have only one match per cell string in S. How I can convert this output into cell arrays of strings in a vectorized form?

+3  A: 

The problem is even worse than you thought. Your output from REGEXP is actually a cell array of cell arrays of cell arrays of strings! Yeah, three levels! The following uses CELLFUN to get rid of the top two levels, leaving just a cell array of strings:

cellArrayOfStrings = cellfun(@(c) c{1},res);

However, you can also change your call to REGEXP to get rid of one level, and then use VERTCAT:

res = regexp(S,'(\d)','tokens','once');  %# Added the 'once' option
cellArrayOfStrings = vertcat(res{:});
gnovice
So easy. :) Thanks a lot!
yuk
@gnovice: You are right about those levels. I think the second solution is better for the speed and clearness. Thanks again.
yuk
@gnovice: I don't know how often I have read the documentation of regexp - yet I haven't noticed the 'once' option. Thanks!
Jonas