Main Content

Access Data in Cell Array

Open Live Script

Basic Indexing

A cell array is a data type with indexed data containers called cells. Each cell can contain any type of data. Cell arrays are often used to hold data from a file that has inconsistent formatting, such as columns that contain both numeric and text data.

For instance, consider a 2-by-3 cell array of mixed data.

C = {'one','two','three'; 
     100,200,rand(3,3)}
C=2×3 cell array
    {'one'}    {'two'}    {'three'   }
    {[100]}    {[200]}    {3x3 double}

Each element is within a cell. If you index into this array using standard parentheses, the result is a subset of the cell array that includes the cells. 

C2 = C(1:2,1:2)
C2=2×2 cell array
    {'one'}    {'two'}
    {[100]}    {[200]}

To read or write the contents within a specific cell, enclose the indices in curly braces.

R = C{2,3}
R = 3×3

    0.8147    0.9134    0.2785
    0.9058    0.6324    0.5469
    0.1270    0.0975    0.9575


C{1,3} = 'longer text in a third location'
C=2×3 cell array
    {'one'}    {'two'}    {'longer text in a third location'}
    {[100]}    {[200]}    {3x3 double                       }

To replace the contents of multiple cells at the same time, use parentheses to refer to the cells and curly braces to define an equivalently sized cell array.

C(1,1:2) = {'first','second'}
C=2×3 cell array
    {'first'}    {'second'}    {'longer text in a third location'}
    {[  100]}    {[   200]}    {3x3 double                       }

Read Data from Multiple Cells

Most of the data processing functions in MATLAB® operate on a rectangular array with a uniform data type. Because cell arrays can contain a mix of types and sizes, you sometimes must extract and combine data from cells before processing that data. This section describes a few common scenarios.

Text in Specific Cells

When the entire cell array or a known subset of cells contains text, you can index and pass the cells directly to any of the text processing functions in MATLAB. For instance, find where the letter t appears in each element of the first row of C.

ts = strfind(C(1,:),'t')
ts=1×3 cell array
    {[5]}    {0x0 double}    {[8 11 18 28]}

Numbers in Specific Cells

The two main ways to process numeric data in a cell array are:

  • Combine the contents of those cells into a single numeric array, and then process that array.

  • Process the individual cells separately.

To combine numeric cells, use the cell2mat function. The arrays in each cell must have compatible sizes for concatenation. For instance, the first two elements of the second row of C are scalar values. Combine them into a 1-by-2 numeric vector.

v = cell2mat(C(2,1:2))
v = 1×2

   100   200

To process individual cells, you can use the cellfun function. When calling cellfun, specify the function to apply to each cell. Use the @ symbol to indicate that it is a function and to create a function handle. For instance, find the length of each of the cells in the second row of C.

len = cellfun(@length,C(2,:))
len = 1×3

     1     1     3

Data in Cells with Unknown Indices

When some of the cells contain data that you want to process, but you do not know the exact indices, you can use one of these options:

  • Find all the elements that meet a certain condition using logical indexing, and then process those elements.

  • Check and process cells one at a time with a for- or while-loop.

For instance, suppose you want to process only the cells that contain character vectors. To take advantage of logical indexing, first use the cellfun function with ischar to find those cells.

idx = cellfun(@ischar,C)
idx = 2x3 logical array

   1   1   1
   0   0   0

Then, use the logical array to index into the cell array, C(idx). The result of the indexing operation is a column vector, which you can pass to a text processing function, such as strlength.

len = strlength(C(idx))
len = 3×1

     5
     6
    31

The other approach is to use a loop to check and process the contents of each cell. For instance, find cells that contain the letter t and combine them into a string array by looping through the cells. Track how many elements the loop adds to the string array in variable n.

n = 0;
for k = 1:numel(C)
    if ischar(C{k}) && contains(C{k},"t")
        n = n + 1;
        txt(n) = string(C{k});
    end
end
txt
txt = 1x2 string
    "first"    "longer text in a third location"

Index into Multiple Cells

If you refer to multiple cells using curly brace indexing, MATLAB returns the contents of the cells as a comma-separated list. For example,

C{1:2,1:2}

is the same as

C{1,1}, C{2,1}, C{1,2}, C{2,2}.

Because each cell can contain a different type of data, you cannot assign this list to a single variable. However, you can assign the list to the same number of variables as cells. 

[v1,v2,v3,v4] = C{1:2,1:2}
v1 = 
'first'

v2 = 
100

v3 = 
'second'

v4 = 
200

If each cell contains the same type of data with compatible sizes, you can create a single variable by applying the array concatenation operator [] to the comma-separated list.

v = [C{2,1:2}]
v = 1×2

   100   200

If the cell contents cannot be concatenated, store results in a new cell array, table, or other heterogeneous container. For instance, convert the numeric data in the second row of C to a table. Use the text data in the first row of C for variable names.

t = cell2table(C(2,:),VariableNames=C(1,:))
t=1×3 table
    first    second    longer text in a third location
    _____    ______    _______________________________

     100      200               {3x3 double}

Index into Arrays Within Cells

If a cell contains an array, you can access specific elements within that array using two levels of indices. First, use curly braces to access the contents of the cell. Then, use the standard indexing syntax for the type of array in that cell.

For example, C{2,3} returns a 3-by-3 matrix of random numbers. Index with parentheses to extract the second row of that matrix.

C{2,3}(2,:)
ans = 1×3

    0.9058    0.6324    0.5469

If the cell contains a cell array, use curly braces for indexing, and if it contains a structure array, use dot notation to refer to specific fields. For instance, consider a cell array that contains a 2-by-1 cell array and a scalar structure with fields f1 and f2.

c = {'A'; ones(3,4)};
s = struct("f1",'B',"f2",ones(5,6)); 
C = {c,s}
C=1×2 cell array
    {2x1 cell}    {1x1 struct}

Extract the arrays of ones from the nested cell array and structure.

A1 = C{1}{2}
A1 = 3×4

     1     1     1     1
     1     1     1     1
     1     1     1     1


A2 = C{2}.f2
A2 = 5×6

     1     1     1     1     1     1
     1     1     1     1     1     1
     1     1     1     1     1     1
     1     1     1     1     1     1
     1     1     1     1     1     1

You can nest any number of cell and structure arrays. Apply the same indexing rules to lower levels in the hierarchy. For instance, these syntaxes are valid when the referenced cells contain the expected cell or structure array.

C{1}{2}{3}

C{4}.f1.f2(1)

C{5}.f3.f4{1}

At any indexing level, if you refer to multiple cells, MATLAB returns a comma-separated list. For details, see Index into Multiple Cells.

See Also

| |

Related Topics