pre-allocating structures
3 visualizaciones (últimos 30 días)
Mostrar comentarios más antiguos
Hello everyone,
I'm trying to pre-allocate memory for a structure defined as follows:
% S : Structure containing geometrical description of polygons.
% S(i) contains all information relative to the i-th shape
% S(i).P(j) gives access to the geometrical description of the j-th
% element of the i-th shape.
% XData : S(i).P(j).x : Vector
% YData : S(i).P(j).y : Vector
% Hole : S(i).P(j).hole : Binary value
I'd like to pre-allocate before the following loop:
for i=1:ne
for j=1:2
S(i).P(j).x = v_proj(f(i,j),1);
S(i).P(j).y = v_proj(f(i,j),2);
S(i).P(j).hole = 0;
end
end
I'd like to have a number of shapes ne.
But I don't really know how to do that. Any help would be more than welcome.
Thank you very much,
% Romain
0 comentarios
Respuesta aceptada
Oleg Komarov
el 1 de Mzo. de 2012
One way to preallocate:
P(1:2) = struct('x',zeros(10,1),'y',zeros(1,2),'hole',false);
ne = 10;
[S(1:ne).P] = deal(P);
Más respuestas (1)
Jan
el 1 de Mzo. de 2012
You can create the last element at first to do an implicite pre-allocation:
for i=ne:-1:1
for j=[2,1]
S(i).P(j).x = v_proj(f(i,j),1);
S(i).P(j).y = v_proj(f(i,j),2);
S(i).P(j).hole = 0;
end
end
2 comentarios
Oleg Komarov
el 1 de Mzo. de 2012
It doesn't work since structures use pointers:
S(20).P(20).x = 1:20;
S : 816.00 B
S(19).P(19).x = 1:20;
S : 1.37 KB
Jan
el 13 de Dic. de 2012
Editada: Jan
el 13 de Dic. de 2012
@Oleg: What does not work and why are the sizes of S interesting here? The code I've posted pre-allocates S in one step and the fields P for each S(:) in one step also. There is no way to pre-allocate all ne elements of P simulataneously, because they are distinct arrays. As far as I can see [S(1:ne).P] = deal(P) creates shared data copies for the contents of all S(:).P, such that writing values later on will consume additional time by creating deep data copies at first. Therefore your pre-allocation is faster only if no data are written and this is rarely needed in real world programs.
Ver también
Categorías
Más información sobre Interpolation en Help Center y File Exchange.
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!