How do I pre-allocate memory for a structure?
이전 댓글 표시
Can anyone help me to fix pre-allocation memory for a structure output?
I have a structure function which have 18 elements, it has two imputs and give results 16 outputs and two inputs.
If I want to pre-allocate the output memory how can I do it.
Here is the code:
phase_in = ?????; % what should be in question marks ( I did zeros(length(v_in),refprop))
for i=1:length(v_in)
phase_in(i) = refprop(v_in(i),T_in,r410a);
end
r410 is my structure function, and v_in and T_in are inputs, I will calculate 100 times to calculate phase_in.
If I skip pre-allocation, the output will give me 1*1000 structure. In this case, I tried to pre-allocate phase_in=zeros(100,18),
but it gave me error'' Conversion to double from struct is not possible.''
Thanks a lot in advance.
댓글 수: 1
By far the fastest and simplest an solution is to allocate the last element first, for example by simply looping over the indices backwards (as Walter Roberson explained in their answer). This topic has been discussed on the MATLAB blogs:
and also in many other Answers threads (with examples) e.g.:
Do NOT use a cell array of structures, this just makes your data storage more complex and less efficient, and means that you cannot use a simple comma-separated list for accessing the structure data:
채택된 답변
추가 답변 (2개)
Walter Roberson
2019년 3월 27일
3 개 추천
Loop backwards from length down to 1 so that the highest offset is assigned to first so it will not need to grow the array afterwards.
댓글 수: 2
Looping backwards is very easy:
for k = numel(v_in):-1:1 % backwards!!
S(k) = ... scalar structure
end
Just make sure that the output structure S does NOT exist in the workspace before the loop.
Walter Roberson
2019년 3월 27일
(Or if it does exist, that it is a compatible structure and no larger than what will be needed.)
Michael
2022년 6월 15일
편집: Steven Lord
2023년 7월 7일
0 개 추천
repmat offers another solution for preallocating structures:
[SL: Removed broken link to code generation documentation, added link to the main repmat function documentation page.]
카테고리
도움말 센터 및 File Exchange에서 Data Type Identification에 대해 자세히 알아보기
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!