How do I remove the empty cells from a vector of cells?

I have a vector of cells which contain strings. Some of the cells in the vector are empty. I want to remove the empty cells from the vector.

Suppose I start with

strs = {'one','','two','three','','','four',''};

and I want to end with

strs = {'one','two','three','four'};

[Rajiv Kumar]

if I have B = {[1 0 0 4],[0 0 0 0],[0 0 1 0],[0 0 2 3]} I want to remove zeros entries like this B = {[1 4],[],[1],[2 3]} How it is possible in cell vector

[Jan]

@Rajiv: This is a completely different question. Please do not high-jack this thread, but open a new one. Thanks.

Answer 1

[Hy]

The built-in function strcmp can compare a character array to a cell array of strings. Matching cells may be removed by setting them equal to the empty array.

 strs(strcmp('',strs)) = [];

Answer 2

[Matt Fig]

Probably the fastest approach:

strs = strs(~cellfun('isempty',strs)) % Call Built-in string

[Jan]

Or the other way around:

strs(cellfun('isempty',strs)) = []

[Jan]

This method is the most efficient solution.

Answer 3

[Ned Gulley]

Here's one way to do it.

strs = {'one','','two','three','','','four',''};

empties = find(cellfun(@isempty,strs)); % identify the empty cells
strs(empties) = []                      % remove the empty cells

This gives

strs = 

      'one'    'two'    'three'    'four'

[Jan]

CELLFUN(@isempty) is remarkably slower than CELLFUN('isempty') as suggested by Matt Fig.

Answer 4

[Michael Katz]

I wanted to do:

strs = setdiff(strs,{''})

but turns out it reorders the output:

strs = 

      'four'    'one'    'three'    'two'

So, I wound up with this:

[~,ix] = setdiff(strs,{''})
strs = strs(sort(ix))

Answer 5

[Bryan White]

For variety:

 cellstr(strvcat(strs))'