# Convert table to struct

17 views (last 30 days)
DavidL88 on 12 Feb 2021
Edited: DavidL88 on 18 Feb 2021
Hi
How can I convert a table to a struct with two layers (a struct within a struct)? Each subject has completed four conditions in a task, and within each of the four conditions there are 100 trials. So there are 400 trials total per subject. Each trial has a delay. I have this data in a table (with 400 rows per subject). Example table attached. I would like to place this data in a struct. First column in the struct should list Subject ID with one row per subject. Second column would have a 1x4 struct per row. This second layer struct within these cells of the second column would contain four rows (for the four Conditions - 11, 12, 21, 22). Each row listing a 1x100 double. These 1x100 double would be the 100 delays per condition per subject.
Example rows from table
T =
Subject Conditions Trial Delay
"Subject111" 11 95 83
"Subject111" 12 96 84
"Subject222" 21 97 67
"Subject333" 22 98 67
The output struct should look something like this;
Struct =
SubjectName Conditions
'Subject111' 1x4 struct
'Subject222' 1x4 struct
'Subject333' 1x4 struct
Within one of 1x4 struct above
delays
1x100 double
1x100 double
1x100 double
1x100 double
Each 1x100 double would have a single row/100 columns listing the delays for that condition for that subject.

DavidL88 on 18 Feb 2021
Edited: DavidL88 on 18 Feb 2021
I figured out a script that does this.
[testID, SubjectNames] = findgroups(t.Subject);
[testID, Events] = findgroups(t.ConditionSample_Whole);
Events = num2cell(Events)
for i = 1:length(SubjectNames)
for i2 = 1:length(Events)
A(i).Subject = SubjectNames(i)
A(i).Events = Events
A(i).Events(i2,2) = {zeros(1,100)}
end
end
for i = 1:length(SubjectNames)
for i2 = 1:length(Events)
b = (ismember(t.Subject, A(i).Subject)) & (ismember(t.Event, A(i).Events{i2}))
idx = find(b)
A(i).Events(i2,2) = {t.Delay(idx)}
end
end