error using the function 'splitapply'
조회 수: 9 (최근 30일)
이전 댓글 표시
Hi! I have a problem about using 'splitapply'.
Using the matrix 'matrix_out_98' the code works, but using the matrix 'matrix_out_12' does not work.
Would anyone be able to help me understand why this?
% matrix_out = importdata("matrix_out_12.mat");
matrix_out = importdata("matrix_out_98.mat");
max_matrix_out = max(matrix_out(:,2));
max_matrix_out_r = floor(max_matrix_out*10)/10;
%Mention the bins to group data in
j = [0 0.1:0.1:max_matrix_out_r Inf];
%Discretize the data
idx2 = discretize(matrix_out(:,2),j);
%Get the sum of the 2nd column according to the groups
out2 = splitapply(@(x) sum(x), matrix_out(:,2), idx2);
Error:
Error using splitapply (line 111)
For N groups, every integer between 1 and N must occur at least once in the vector of group numbers.
댓글 수: 2
Dyuman Joshi
2023년 11월 30일
이동: Dyuman Joshi
2023년 11월 30일
accumarray will be a better fit here.
matrix_out_12 = importdata("matrix_out_12.mat");
matrix_out_98 = importdata("matrix_out_98.mat");
out_12 = fun(matrix_out_12)
out_98 = fun(matrix_out_98)
function out = fun(matrix_out);
max_matrix_out = max(matrix_out(:,2));
max_matrix_out_r = floor(max_matrix_out*10)/10;
%Mention the bins to group data in
j = [0 0.1:0.1:max_matrix_out_r Inf];
%Discretize the data
idx = discretize(matrix_out(:,2),j);
%Get the sum of the 2nd column according to the groups
out = accumarray(idx, matrix_out(:,2), [], @sum);
end
Dyuman Joshi
2023년 11월 30일
이동: Dyuman Joshi
2023년 11월 30일
splitapply, as the error message states, requires there to be atleast 1 data element in every bin/group in which the data is discretized.
There's no such requirement with accumarray(), where it fills the gap automatically - with the default value of 0, or manually as the value specified (for example NaN)
채택된 답변
Stephen23
2023년 11월 30일
편집: Stephen23
2023년 11월 30일
S = load('matrix_out_12.mat') % LOAD is better than IMPORTDATA
max_matrix_out = max(S.matrix_out(:,2));
max_matrix_out_r = floor(max_matrix_out*10)/10; % use ROUND(..,1)
j = [0:0.1:max_matrix_out_r,Inf];
idx2 = discretize(S.matrix_out(:,2),j);
Lets check if all of those groups exist in your data:
ismember(1:max(idx2),idx2) % Nope, some don't:
SPLITAPPLY cannot cope with that: it applies the function to every group from 1 to N, it does not know how to apply your function to data if that data simply does not exist.
Use ACCUMARRAY instead, which in contrast will happily replace undefined outputs with a fill value (by default zero, but you can also set the fill value to whatever you want):
out = accumarray(idx2,S.matrix_out(:,2),[max(idx2),1],@sum)
out(end-6:end)
For example, using NaN as the fill value:
out = accumarray(idx2,S.matrix_out(:,2),[max(idx2),1],@sum, NaN)
out(end-6:end)
댓글 수: 7
Stephen23
2023년 11월 30일
편집: Stephen23
2023년 11월 30일
"'out1' consists of {A×1 double}, should be {A×2 double}"
So you want both columns. Here is an approach using ARRAYFUN:
load("matrix_out_98.mat")
load("idx1_98.mat")
out = arrayfun(@(n)matrix_out(n==idx1,:),1:max(idx1),'Uni',0)
Checking a few cells:
out{5}
out{9}
Note that often splitting up data make it harder to work with.
추가 답변 (0개)
참고 항목
카테고리
Help Center 및 File Exchange에서 Statistics and Machine Learning Toolbox에 대해 자세히 알아보기
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!