CODEWITHSUNDEEP
pythonindexingbooleanjuliascientific-computing
quics-ilver
quics-ilver

Reputation: 119

Obtaining columns of a matrix satisfying multiple booleans in Julia

I am trying to obtain only those columns of a matrix that satisfy two or more boolean conditions. More specifically, I am trying to find only those eigenvectors of a matrix based on certain constraints on the eigenvalues.

In python3.x I would do :

def get_special_vectors(A:np.ndarray,thresold1:float,thresold2:float) -> np.ndarray:
   eigval, eigvec = np.linalg.eig(A)
   eigvec = eigvec[:,np.array(eigval >= threshold1) & np.array(eigval <= thresold2)]
return eigvec

However in Julia, I am following this post but I seem to be messing up the AND (&) operator. I'm trying the following that results in a MethodError:

eigvec = eigvec[:,vec(evalA .< 1.0) & vec(evalA .> 0)]

I'll be glad if someone can share any useful suggestions or any kind of help. Thanks in advance!

Upvotes: 3

Views: 56

Answers (2)

MarcMush
MarcMush

Reputation: 1488

you should also broadcast the &:

vec(evalA .< 1.0) .& vec(evalA .> 0)

I can't say for sure if it will solve your problem since you didn't give a complete example

You also might want to look at eachrow and eachcol

Upvotes: 3

Bogumił Kamiński
Bogumił Kamiński

Reputation: 69879

You need to broadcast also & so write .& instead like this:

eigvec[:,vec(evalA .< 1.0) .& vec(evalA .> 0)]

however in this case the following should also just work:

eigvec[:,vec(0 .< evalA .< 1.0)]

(I do not see what evalA is, so it is hard to tell if vec is actually required in your code - I assume that it is required so I left it)

Upvotes: 2

Related Questions