CODEWITHSUNDEEP
matlabsplit
ANieder
ANieder

Reputation: 223

Automatic split of cell array according to a column values into variable number of new cell arrays

Some time ago I asked this question: Automatic split of character matrix according to a column values into variable number of new dataframes

For that question the function split was the perfect answer, now, I am triying to do the same in Matlab and cannot find any function that does the same.

So, its simple, my question is: Is there any function that does the same as the R split function in Matlab? and if not, how could I do exactly what was asked in the linked question in Matlab?

Upvotes: 1

Views: 4384

Answers (2)

Andy Campbell
Andy Campbell

Reputation: 2187

I think that what you need is the new table class R2013b:

>> x = {'Hi', 'Med', 'Hi', 'Low'; 'A', 'D', 'A', 'C'; '8', '3', '9', '9'; '1', '1', '1', '2'}';
>> t = cell2table(x)

t = 

     x1      x2     x3     x4 
    _____    ___    ___    ___

    'Hi'     'A'    '8'    '1'
    'Med'    'D'    '3'    '1'
    'Hi'     'A'    '9'    '1'
    'Low'    'C'    '9'    '2'

>> slice = t(strcmp(t.x2,'A'),:)

slice = 

     x1     x2     x3     x4 
    ____    ___    ___    ___

    'Hi'    'A'    '8'    '1'
    'Hi'    'A'    '9'    '1'

>>

For more information, take a look at the table class here

If you'd like to do this for all unique values of the table you can leverage cellfun:

>> slices = cellfun(@(value) t(strcmp(t.x2,value),:), unique(t.x2),'UniformOutput',false)

slices = 

    [2x4 table]
    [1x4 table]
    [1x4 table]

>> slices{:}

ans = 

     x1     x2     x3     x4 
    ____    ___    ___    ___

    'Hi'    'A'    '8'    '1'
    'Hi'    'A'    '9'    '1'


ans = 

     x1      x2     x3     x4 
    _____    ___    ___    ___

    'Low'    'C'    '9'    '2'


ans = 

     x1      x2     x3     x4 
    _____    ___    ___    ___

    'Med'    'D'    '3'    '1'

>>

In this case you are passing every unique element of t.x2 into this function. Uniform output must be false so that cellfun does not try to concatenate the tables together but rather puts each slice of the table into its own cell.

Upvotes: 1

StrongBad
StrongBad

Reputation: 879

The unique and strcmp functions works on cell arrays. What about something like

x = {'Hi', 'Med', 'Hi', 'Low'; 'A', 'D', 'A', 'C'; '8', '3', '9', '9'; '1', '1', '1', '2'}';

for ii = unique(x(:, 3))'
    x(find(strcmp(x(:, 3), ii)), :)
end

While it may seems silly, the index array for the for loop must be a row so in unique(x(:, 3))' the transpose operator is critical. If done correctly ii is a scalar. Otherwise you will get the error

Error using strcmp Inputs must be the same size or either one can be a scalar

Upvotes: 1

Related Questions