Accessing a cell in a 2D grid

Hi all,

I have a 2D grid as shown below where each cell represents a pair of and where and . I would like to construct a new index that combines these two indices to run over all of the cells as a one continous number.

Any help would be appreciated.

Thanks.

Answer 1

[Cris LaPierre]

Consider using linear indexing. You can find an explanation towards the middle of this page. It is also explained here. If you need to go back and forth between a linear index and row/column indexing, you can use the functions ind2sub and sub2ind.

[dpb]

You should also consider whether you actually need to address the individual cells directly by index or whether couldn't use vectorized expressions.

We don't know the perceived need/use for this so can't give specifics...

[Lama Hamadeh]

Many thanks for your reply. However, I still don't have an answer for my question. Probably it's my fault I didn't explain better, but what I need is a mathematical formula that depends on (i) and (j) that represents a given cell. Hopefully the below sketch can describe what I have in mind:

[dpb]

" what I need is a mathematical formula that depends on (i) and (j) that represents a given cell. "

Well, that's pretty easy to derive from the ordering chosen and the size of the array. Look at the pattern in the numbers you've written down--it's a linear expression in N and M.

Why nobody has given it to you is because in MATLAB there's probably a much more efficient way to code whatever it is you're after than by such an expression so we're trying to guide you towards that solution instead.

NB: Your ordering above is reversed from MATLAB storage order in going from LLH corner up rather than ULH corner down.

[dpb]

BTW, the "formula" in MATLAB is

k=ind2sub(size(A),i,j);

which can be written as

k=ind2sub([M,N],i,j);

if you have the dimensions instead.

>> fnNDX=@(SZ,i,j) SZ(1).*(i-1)+j
fnNDX =
  function_handle with value:
    @(SZ,i,j)SZ(1).*(i-1)+j
>> [X,Y]=meshgrid([1:4].',[1:4].');
>> fnNDX([4,4],X,Y)
ans =
     1     5     9    13
     2     6    10    14
     3     7    11    15
     4     8    12    16
>> 

[Lama Hamadeh]

That is amazing! Many thanks for that!

[dpb]

Yet again, however, I'd almost wager it isn't the most fruitful approach to the problem...