How can I find index of element in array?

I know that I have a number 5 as an element in array X, but I do not know its index. Does MATLAB have a built-in function similar to Python's "index" method for finding the index of an element in an array?

Answer 1

[James Tursa]

To find the index of a specific integer value (without roundoff error) in an array of integers, use the "find" function and == operator. For example, find the index of an element equal to 5 in a 1-by-11 vector of integers. 

x = 0:1:10; 
k = find(x==5) 

To find a numeric value in an array of floating-point numbers, use a tolerance value based on your data. Otherwise, the result is sometimes an empty matrix due to floating-point roundoff errors. For example, find the index of an element equal to 0.5 within a roundoff error of 1e-6. 

y = 0:0.1:1; 
k = find(abs(y-0.5) < 1e-6) 

[Stephen23]

Often logical indexing is more efficient, so you might only need this:

idx = X==5;

[Scott MacKenzie]

Bear in mind that if the number occurs more than once in the vector, the result returned is a vector containing the indices of all occurrences. If you want the index of just the first occurrence of the number, insert 1 as the second argument in find:

>> x = [3 4 5 6 4 8]
x =
     3     4     5     6     4     8
     
>> result = find(x==4)
result =
     2     5

>> result = find(x==4, 1)
result =
     2

[Walter Roberson]

Henrik Wassertheurer comments to James Tursa

Doesn't answer the question

[Image Analyst]

Henrik, yes it does, as I understand the question. Why do you say it doesn't?

[Walter Roberson]

@Henrik Wassertheurer

What difficulty do you find with James' answer? He showed the find() function, which is the function defined to locate the places where a condition occurs.

If you need to have the exact question answered more clearly, "but did Matlab doesn't have build-in similar function?" then the answer to that is "NO, MATLAB does not have a built-in function in which you can provide only the array name and the value, and MATLAB will return all the indices of the value in the array."

Note: if you only need to know the first location, then you can also use

[~, result] = ismember(5, x)

result will be 0 if 5 is not present in x.

[Ehsan Partovi]

The function find() is useful as far as matrices (2-D tensors) are concerned. I cannot, however, find a useful function for nd-arrays where, for instance, the index could be an array on its own. See example below:

M = reshape(1:24, [2,3,4]);

indices = index_finder(M==20); % indices = vector of indices

It would be very useful if there was a function which worked for tensors of any dimensionality.

[Jesse Ivers]

@Ehsan Partovi I couldn't agree with you more; this is a problem I seem to run into often, and here is my solution:

% Example ND-array
arr = reshape([1:6000], [5 5 10 4 6]);
numberOfInterest = 99;

% Get the linear index of the 
linearIndex = find(arr==numberOfInterest);
% Convert linear index to subscript
[row, col, depth, channel, time] = ind2sub(size(arr), linearIndex)

row = 4

col = 5

depth = 4

channel = 1

time = 1

The only drawbacks are the reuirement that you know how many dimensions. YOu can get around this with CSLs like so:

% Use CSL to get all the outputs
[idicies{1:ndims(arr)}] = ind2sub(size(arr), linearIndex)

idicies = 1×5 cell array
    {[4]}    {[5]}    {[4]}    {[1]}    {[1]}