# Find indexs of minimum in 4-D matrix.

15 views (last 30 days)
Clément Métayer on 10 May 2020
Commented: Steven Lord on 10 May 2020
Hi im trying to find the indexs of the minimum in a 4-D "matrix" but the usual command:
[Minimum,Index]=min(Expression,[],'all') does not work.
(here "Expression" is my function of 4 variables )
The error message tells me : Error using min
Specify 'linear' option when returning two outputs and the dimension is 'all' or a vector.
But when i specify the linear option it returns me the index 11755218 and i don't understand what it means ?
Note: the ndgrid of my function is :[phi1,phi2,phi3,phi4]=ndgrid(0:0.1:2*pi,0:0.1:2*pi,0:0.1:2*pi,0:0.1:2*pi);
Thx to the great saouls who can help me.
Steven Lord on 10 May 2020
Walter has showed you how to get what you're looking for. But to answer your question (what 11755218 means) that is the linear index to the element whose value is returned as Minimum. See the "Indexing with a Single Index" section on this documentation page for an explanation of linear indexing.
A = reshape(1:24, [4 2 3])
A(15)
A(3, 2, 2)
Element (3, 2, 2) of A is also element 15 of A.

Walter Roberson on 10 May 2020
Edited: Walter Roberson on 10 May 2020
[Minimum, Index] = min(Expression, [], 'all', 'linear')
[idx{1:ndims(Expression)}] = ind2sub(size(Expression), Index);
[idx{:}]
Though perhaps what you want is instead
[Minimum, Index] = min(Expression, [], 'all', 'linear')
[phi1(Index), phi2(Index), phi3(Index), phi4(Index)]
Clément Métayer on 10 May 2020
Yes the second is the rgiht one that i was looking for.
Thx walter

Gifari Zulkarnaen on 10 May 2020
Maybe linear means the index ordered in from high dimension to one dimension. For example, for 2D matrix with 2 rows and 3 columns:
example2D = [3 2 1; 4 5 6];
[minValue,minIndex] = min(example2D,[],'all','linear');
Here minIndex = 5 means row 1 and column 3 (2x2+1).
For your case also, Expression(Index) is equal to Minimum.