I have a vector like:
a = [1,2,3,4,5,6...,n]
and I would like to obtain a new vector like this:
a_new = [1,0,0,2,0,0,3,0,0,4,0,0,5,0,0,6,...,0,0,n]
where a fixed number of zeros (2 in the above example) are inserted between the non-zero elements. If I choose zero_p=3, the new vector would be:
a_new = [1,0,0,0,2,0,0,0,3,0,0,0,4,0,0,0,5,0,0,0,6,...,0,0,0,n]
etc.
How can I do this?
Try this:
zero_p=3;
a_new=zeros(1, (zero_p+1)*length(a)-zero_p);
a_new(1:(zero_p+1):end)=a;
(Untested, but should hopefully work.)
There's a few ways I can think of:
Kronecker product
The kronecker product is excellently suited for this.
In Matlab, kron is what you're looking for:
a = 1:4;
a = kron(a, [1 0 0])
ans =
1 0 0 2 0 0 3 0 0 4 0 0
or, generalized,
a = 1:4;
zero_p = 3;
b = [1 zeros(1,zero_p-1)];
a = kron(a, b)
ans =
1 0 0 2 0 0 3 0 0 4 0 0
If you want to have it end with a non-zero element, you have to do one additional step:
a = a(1:end-zero_p);
Or, if you like one-liners, the whole thing can be done like this:
a = 1:4;
zero_p = 3;
a = [kron(a(1:end-1), [1 zeros(1,zero_p-1)]), a(end)]
ans =
1 0 0 2 0 0 3 0 0 4
Zero padding
Probably the simplest method and best performance:
a = 1:4;
zero_p = 3;
a = [a; zeros(zero_p, size(a, 2))];
a = a(1:end-zero_p);
Matrix multiplication
Also simple, readable and great performance, although it might be overkill for many situations other than this particular scenario:
a = 1:4;
b = [1; zeros(zero_p, 1)];
a = b*a;
a = a(1:end-zero_p);
x = [1 2 3 4 5];
upsample(x,3)
o/p: 1 0 0 2 0 0 3 0 0 4 0 0 5 0 0
Cheers!!
Related
I need a function that creates a checker board matrix with M rows and N columns of P*Q rectangles. I modified the third solution from here to get that:
function [I] = mycheckerboard(M, N, P, Q)
nr = M*P;
nc = N*Q;
i = floor(mod((0:(nc-1))/Q, 2));
j = floor(mod((0:(nr-1))/P, 2))';
r = repmat(i, [nr 1]);
c = repmat(j, [1 nc]);
I = xor(r, c);
it works with no problem:
I=mycheckerboard(2, 3, 4, 3)
I =
0 0 0 1 1 1 0 0 0
0 0 0 1 1 1 0 0 0
0 0 0 1 1 1 0 0 0
0 0 0 1 1 1 0 0 0
1 1 1 0 0 0 1 1 1
1 1 1 0 0 0 1 1 1
1 1 1 0 0 0 1 1 1
1 1 1 0 0 0 1 1 1
But it's not fast enough since there are lots of calls of this function in a single run. Is there a faster way to get the result? How can I remove floating point divisions and/or calls of the floor function?
Your code is fairly fast for small matrices, but becomes less so as the dimensions get larger. Here's a one-liner using bsxfun and imresize (requires Image Processing toolbox that most have):
m = 2;
n = 3;
p = 4;
q = 3;
I = imresize(bsxfun(#xor, mod(1:m, 2).', mod(1:n, 2)), [p*m q*n], 'nearest')
Or, inspired by #AndrasDeak's use of kron, this is faster with R2015b:
I = kron(bsxfun(#xor, mod(1:m, 2).', mod(1:n, 2)), ones(p, q))
For a small bit more speed, the underlying code for kron can be simplified by taking advantage of the structure of the problem:
A = bsxfun(#xor, mod(1:m, 2).', mod(1:n, 2));
A = permute(A, [3 1 4 2]);
B = ones(q, 1, p);
I = reshape(bsxfun(#times, A, B), [m*n p*q]);
or as one (long) line:
I = reshape(bsxfun(#times, permute(bsxfun(#xor, mod(1:m, 2).', mod(1:n, 2)), [3 1 4 2]), ones(q, 1, p)), [m*n p*q]);
I suggest first creating a binary matrix for the checkerboard's fields, then using the built-in kron to blow it up to the necessary size:
M = 2;
N = 3;
P = 4;
Q = 3;
[iM,iN] = meshgrid(1:M,1:N);
A = zeros(M,N);
A(mod(iM.'+iN.',2)==1) = 1;
board = kron(A,ones(P,Q))
I need to check if a matrix is a diagonal block matrix or not. Is there any easy way to check it? Especially, that would be perfect if a simple function such as isdiag()exist.
A function isdiag actually exists for R2014a and later. It works for square and non-square matrices. For earlier versions of Matlab see an alternative below.
Here is a quick demo:
A = [1 0 0;
0 2 0;
0 0 3];
isdiag(A)
B = [1 0 0;
0 2 3;
0 0 4];
isdiag(B)
C = [1 0 0 0;
0 2 0 0;
0 0 3 0];
isdiag(C)
D = [1 0 0 0;
0 2 0 3;
0 0 4 0];
isdiag(D)
... and we get:
ans =
1
ans =
0
ans =
1
ans =
0
In R2014a and later, you can as well use isbanded with a lower and upper bandwidth of 0. This will give you exactly the same result:
isbanded(M,0,0)
For versions before R2014a, you can use a combination of find and all:
[t1,t2] = find(M);
res = all(t1==t2)
I'd like to create a 4x12 matrix which is very similar to a upper triangle matrix, it looks like this:
1 1 1 1 1 1 1 1 1 1 1 1
0 0 0 1 1 1 1 1 1 1 1 1
0 0 0 0 0 0 1 1 1 1 1 1
0 0 0 0 0 0 0 0 0 1 1 1
So my question is. What is the most efficient way to create it? no loops, no cellfun. Thanks.
One vectorized approach -
nrows = 4;
ncols = 12;
row_idx = repmat(1:nrows,ncols/nrows,1)
out = bsxfun(#le,[1:nrows]',row_idx(:).')
The Matlab R2015a and later approach using the newly introduced repelem:
n = 4;
m = 3;
out = repelem(triu(ones(n)),1,m);
out =
1 1 1 1 1 1 1 1 1 1 1 1
0 0 0 1 1 1 1 1 1 1 1 1
0 0 0 0 0 0 1 1 1 1 1 1
0 0 0 0 0 0 0 0 0 1 1 1
It even seems faster than the bsxfun approach, though I can't believe this ;)
Benchmark
Unfortunately I couldn't consider andrew's solution as it is not complete and I didn't got it totally.
function [t] = bench()
n = 4;
m = 12;
t = zeros(3,15);
for ii = 1:15
fcns = {
#() thewaywewalk(ii*n,ii*m);
#() Divakar(ii*n,ii*m);
#() LuisMendo(ii*n,ii*m);
};
% timeit
for jj = 1:100;
t(:,ii) = t(:,ii) + cellfun(#timeit, fcns);
end
end
plot(1:15,t(1,:)); hold on;
plot(1:15,t(2,:)); hold on;
plot(1:15,t(3,:)); hold on;
xlabel('Matrix size: n = x*4, m = x*12')
ylabel('timing')
legend({'thewaywewalk','Divakar','Luis Mendo'},'location','northwest')
end
function Z = thewaywewalk(n,m)
Z = repelem(triu(ones(n)),1,m/n);
end
function Z = Divakar(n,m)
row_idx = repmat(1:n,m/n,1);
Z = bsxfun(#le,[1:n]',row_idx(:).');
end
function Z = LuisMendo(n,m)
Z = reshape(repmat(permute(triu(ones(n,n)), [1 3 2]), [1 m/n 1]), [n m]);
end
First bottomline - small matrices:
The new repelem does a very good job, but also the reshape(repmat(permute... does not disappoint. The bsxfun approach stays a little behind for some medium-sized matrices, before it becomes the leader for large matrices:
Second bottomline - big matrices:
As predicted by Divakar, bsxfun is the fastest for larger matrices, actually as expected as bsxfun is always the fastest! Interesting that the other two align perfectly, on could guess they almost work the same internally.
Create an upper triangular matrix of ones, permute second and third dimensions, repeat along second dimension, and reshape into desired shape:
m = 4;
n = 12;
result = reshape(repmat(permute(triu(ones(m,m)), [1 3 2]), [1 n/m 1]), [m n]);
depending on your matlab version
m = 4;
n = 12;
dec2bin(bitshift(num,-1*[0:n/m:n-1])) %this prints out a string
these should be logical arrays (i don't have either of these so I cant test it)
decimalToBinaryVector(bitshift(num,-1*[0:n/m:n-1]))
de2bi(bitshift(num,-1*[0:n/m:n-1]))
I have a matrix A
A = [0 0 0 0 1; 0 0 0 0 2; 0 1 2 3 4];
and I would like to randomly permute the elements within each row. For example, matrix A2
A2 = [1 0 0 0 0; 0 0 0 2 0; 4 1 3 2 0]; % example of desired output
I can do this with a vector:
Av = [0 1 2 3 4];
Bv = Av(randperm(5));
But I am unsure how to do this a row at time for a matrix and to only permute the elements within a given row. Is this possible to do? I could construct a matrix from many permuted vectors, but I would prefer not to do it this way.
Thanks.
You can use sort on a random array of any size (which is what randperm does). After that, all you need to do is some index-trickery to properly reshuffle the array
A = [0 0 0 0 1; 0 0 0 0 2; 0 1 2 3 4];
[nRows,nCols] = size(A);
[~,idx] = sort(rand(nRows,nCols),2);
%# convert column indices into linear indices
idx = (idx-1)*nRows + ndgrid(1:nRows,1:nCols);
%# rearrange A
B = A;
B(:) = B(idx)
B =
0 0 1 0 0
0 2 0 0 0
2 1 3 4 0
I need a matrix of nxn, where the first pxp of it contains ones and rest are zeros. I can do it with traversing the cells, so I'm not asking a way to do it. I'm looking for "the MATLAB way" to do it, using built-in functions and avoiding loops etc.
To be more clear;
let n=4 and p=2,
then the expected result is:
1 1 0 0
1 1 0 0
0 0 0 0
0 0 0 0
There are possibly more than one elegant solution to do it, so I will accept the answer with the shortest and most readable one.
P.S. The question title looks a bit irrelevant: I put that title because my initial approach would be creating a pxp matrix with ones, then expanding it to nxn with zeros.
The answer is creating a matrix of zeroes, and then setting part of it to 1 using indexing:
For example:
n = 4;
p = 2;
x = zeros(n,n);
x(1:p,1:p) = 1;
If you insist on expanding, you can use:
padarray( zeros(p,p)+1 , [n-p n-p], 0, 'post')
Another way to expand the matrix with zeros:
>> p = 2; n = 4;
>> M = ones(p,p)
M =
1 1
1 1
>> M(n,n) = 0
M =
1 1 0 0
1 1 0 0
0 0 0 0
0 0 0 0
You can create the matrix easily by concatenating horizontally and vertically:
n = 4;
p = 2;
MyMatrix = [ ones(p), zeros(p, n-p); zeros(n-p, n) ];
>> p = 2; n = 4;
>> a = [ones(p, 1); zeros(n - p, 1)]
a =
1
1
0
0
>> A = a * a'
A =
1 1 0 0
1 1 0 0
0 0 0 0
0 0 0 0