MATLAB Answers

loop based on a condition in another column

2 views (last 30 days)
Simon Jurgies
Simon Jurgies on 18 Mar 2021
Edited: Seth Furman on 18 Mar 2021
I'm working with a 694672x4 table containing the PermNo., date, CUSIP and stock price of a multitude of shares. The data is sorted such that the full time series of a stock is listed below the full time series of the following stock and so forth. (Pls see photo below). I am trying to calculate the daily returns of each stock individually, however I struggle with the way the data is structured. As you can see, column 3 contains the CUSIP which is an individual identifier to every stock in the sample. I am now trying to tell MatLab to calculate the daily returns (using (diff(log(X))))for all rows with the same identifier and start from scratch once the identifier changes. Unfortunately, I cannot provide any initial code as I don't know where to start...
Thank you in advance!
  4 Comments
Walter Roberson
Walter Roberson on 18 Mar 2021
Probably you will need
func = @(x) {diff(log(x))}
That is because you will be processing vectors, and diff() of a vector is a vector (one shorter), but splitapply() requires that you return a scalar.

Sign in to comment.

Accepted Answer

Seth Furman
Seth Furman on 18 Mar 2021
Edited: Seth Furman on 18 Mar 2021
To add to Walter's answer, you might also consider groupsummary.
>> rng default
>> t = table([1;1;1;2;2;2;2],rand(7,1),'VariableNames',{'Group','Data'})
t =
7×2 table
Group Data
_____ _______
1 0.81472
1 0.90579
1 0.12699
2 0.91338
2 0.63236
2 0.09754
2 0.2785
>> groupsummary(t,'Group',@(x){diff(log(x))})
ans =
2×3 table
Group GroupCount fun1_Data
_____ __________ ____________
1 3 {2×1 double}
2 4 {3×1 double}

More Answers (0)

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by