Concatenate Columns of Cells

3 views (last 30 days)
Chameleon17
Chameleon17 on 2 Jun 2015
Commented: Stephen23 on 22 Dec 2021
Hi,
I have asked this question earlier, but I think I have though of how to better phrase it.
Columns 1 through 5
'01/04/2004' [ 0] [ 0] [ 0] [ 0]
'01/04/2004' [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] '05/04/2004'
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] '04/04/2004' [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] '04/04/2004' [ 0]
[ 0] '02/04/2004' [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] '05/04/2004'
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] '05/04/2004'
[ 0] [ 0] [ 0] '04/04/2004' [ 0]
[ 0] [ 0] [ 0] '04/04/2004' [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
'01/04/2004' [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] '03/04/2004' [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
'01/04/2004' [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] '03/04/2004' [ 0] [ 0]
[ 0] [ 0] '03/04/2004' [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] '05/04/2004'
There is only ever one date per row, and I want to concatenate them into one column, removing all the zeros.
I think it is difficult because of the cell/date format that I have going on. I've managed to make one very long vector of it all, and get rid of the zeros but lose the order they are in.
Any thoughts would be very much appreciated! Or any reading, or hints even...
Thanks!

Answers (3)

Jan Orwat
Jan Orwat on 2 Jun 2015
Hello, I can see that date is of type string in your cell array and non-date fields are zeros (non-string). Assuming that in every row there is just one date you can get those like that:
c = yourCell.'; % rows becomes columns, columns becomes rows - this is for MATLAB indexing purposes
datesOnly = c(cellfun(@isstr,c));
This method will delete those rows which doesn't contain dates at all. It won't work properly if row contain more than one date because it gets all dates from array row by row.
  2 Comments
Jan Orwat
Jan Orwat on 2 Jun 2015
Hope this will help:
>> a = {
[ 0] [ 0] '20/7/2014'
[ 0] '23/7/2014' [ 0]
'26/7/2014' [ 0] [ 0]
[ 0] [ 0] [ 0]
'21/7/2014' [ 0] [ 0]};
>> [r,c] = find(cellfun(@isstr,a));
>> dates = accumarray(r,r+(c-1)*size(a,1),[],@(x)a(x),{'missing'})
dates =
'20/7/2014'
'23/7/2014'
'26/7/2014'
'missing'
'21/7/2014'
>> % or:
>> dates = accumarray(r,r+(c-1)*size(a,1),[],@(x)a(x),{0})
dates =
'20/7/2014'
'23/7/2014'
'26/7/2014'
[0]
'21/7/2014'

Sign in to comment.


Thomas Koelen
Thomas Koelen on 2 Jun 2015
I think this does what you want :)
a={'20/7/2014' 0 0; 0 '23/7/2014' 0; 0 0 '26/7/2014'};
j=1;
for i=1:size(a,1)*size(a,2)
if a{i}~=0
x(j)=i;
j=j+1;
end
end
for k=1:j-1
anew{k,1}=a{x(k)};
end
anew
a =
'20/7/2014' [ 0] [ 0]
[ 0] '23/7/2014' [ 0]
[ 0] [ 0] '26/7/2014'
anew =
'20/7/2014'
'23/7/2014'
'26/7/2014'
  1 Comment
Jan Orwat
Jan Orwat on 2 Jun 2015
Hi Thomas, in your example, when you change a:
>> a={0 0 '20/7/2014'; 0 '23/7/2014' 0; '26/7/2014' 0 0}
a =
[ 0] [ 0] '20/7/2014'
[ 0] '23/7/2014' [ 0]
'26/7/2014' [ 0] [ 0]
it gives:
>> anew
anew =
'26/7/2014'
'23/7/2014'
'20/7/2014'
so it looks like it does not work.

Sign in to comment.


Andrei Bobrov
Andrei Bobrov on 2 Jun 2015
d = { '01/04/2004' [ 0] [ 0] [ 0] [ 0]
'01/04/2004' [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] '05/04/2004'
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] '04/04/2004' [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] '04/04/2004' [ 0]
[ 0] '02/04/2004' [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] '05/04/2004'
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] '05/04/2004'
[ 0] [ 0] [ 0] '04/04/2004' [ 0]
[ 0] [ 0] [ 0] '04/04/2004' [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
'01/04/2004' [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] '03/04/2004' [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
'01/04/2004' [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] '03/04/2004' [ 0] [ 0]
[ 0] [ 0] '03/04/2004' [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] [ 0]
[ 0] [ 0] [ 0] [ 0] '05/04/2004'};
[ii,jj] = find( cellfun('ischar',d));
k = sortrows([ii,jj],1);
out = d(sub2ind(size(d),k(:,1),k(:,2)));
  2 Comments
Stephen23
Stephen23 on 22 Dec 2021
The supported backward compatibility** options are described here:
and the correct syntax is:
cellfun('isclass',C,'char')
** It is worth nothing that these are more efficient than supplying a function handle.

Sign in to comment.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!