Get max and min of each row in a table and create a new table with those values
5 vues (au cours des 30 derniers jours)
Afficher commentaires plus anciens
i have a table with returns of 10 different portfolios for each month. Columns are the names of the portfolios and each row corresponds to one month. I want to rebuild a momentum portfolio, so in the first row(first month) I look at the 3 best performers and 3 worst performers. After determining which Portfolio performed best/worst, I will invest in it for 1 month. This means that I need the return of month 2 (row 2) for the best/worst performers of row 1. In row 2 (month 2) I look again at the best/worst performers and I need their return from row 3 (month 3) and so on...Those values should be in a new table so I can plot it later.
Screenshot of first 5 rows attached. the first column is the date, I want to ignore that one.
I'm very new to Matlab and tried to do it on my own but did not succeed at all.
My initial thought was to go with
[maxVals,maxLocs] = maxk(A{:,vartype('numeric')},2,2)
[minVals,minLocs] = mink(A{:,vartype('numeric')},2,2)
But I don't know how to get the corresponding return of the next month associated with the best/worst performers of the previous month.
Thank you all!
0 commentaires
Réponses (1)
Andrei Bobrov
le 9 Avr 2020
Modifié(e) : Andrei Bobrov
le 9 Avr 2020
Let A - your table.
[~,i] = sort(A{1,2:end});
out = A(:,[1,1+i([end:-1:end-2,3:-1:1])]);
2 commentaires
Andrei Bobrov
le 10 Avr 2020
Modifié(e) : Andrei Bobrov
le 10 Avr 2020
Maybe it's
T = readtable('Path\your\file\Screenshot 2020-04-09 at 11.34.40.xlsx',...
'ReadVariableNames',1);
T.Date = datetime(T.Date,'InputFormat','yyddMM');
N = string(T.Properties.VariableNames(2:end));
[~,i] = sort(T{:,2:end},2,'descend');
out = [T(:,1),array2table(N(i(:,[1:3,end-2:end])))];
Voir également
Catégories
En savoir plus sur Matrices and Arrays dans Help Center et File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!