MATLAB Answers

splitapply with own created function

조회 수: 3(최근 30일)
I have 'nr' tables containing the following: date/time/Qdot/temperature
i would like to apply a function to this tables more specific i would like to apply a function on data grouped by day (date)
herefor i've tried to use the function splitapply where the function is of my own creating
for j=2:24
s = (j-1); %sizetimestep - 1stap
for i=1:nr
[G,days] = findgroups(DayT{i}.date);
max_Qdot = splitapply(@maxperinterval,DayT{i}.Qdot,s,G);
Mean_T = splitapply(@mean,DayT{i}.Temperature,G);
Max_tabel{j,i}=table(days,max_Qdot,Mean_T,'VarbiableNames',rownames);
end
end
But when i tried to run this i get an error saying the following:
"The data variables must have the same number of rows as the vector of group numbers. The group number vector
has 5736 row(s), and data variable 2 has 1 row(s)."
My own function needs the input of variable s.
Is this even possible to use this function with splitapply?

  댓글 수: 0

로그인 to comment.

채택된 답변

Rik
Rik 1 Apr 2020
편집: Rik 1 Apr 2020
You can probably get around this by wrapping your function in an anonymous function:
fun=@(input1)maxperinterval(input1,s);
max_Qdot = splitapply(fun,DayT{i}.Qdot,G);
Now the function splitapply is looking at only requires 1 input variable while your function will receive both. Note that you will need to generate fun every time you update s.

  댓글 수: 1

Senne Van Minnebruggen
Thanks! This solved my problem.

로그인 to comment.

추가 답변(1개)

Ameer Hamza
Ameer Hamza 1 Apr 2020
As the error indicate. The error is related to the orientation of matrices. Change it like this
max_Qdot = splitapply(@maxperinterval,DayT{i}.Qdot,s.',G); % use transpose of s so that it become column matrix

  댓글 수: 0

로그인 to comment.

이 질문에 답변하려면 로그인을(를) 수행하십시오.

태그


Translated by