How can I concatenate vectors based on a grouping label contained by a separate variable?

Hello all,

I have 16 test subjects. Each subject underwent a different number of tests. For each test, I have a results vector. I would like to concatenate all the results vectors of the same subject.

Let met illustrate this with psuedo code.

Imagine I have the following test labels as strings, and in a separate variable, I have their test results vectors, each of which may be of different length. The indices of these two seperate variables match up, so that the nth test label corresponds to the nth results vector:

Test labels Results vector

'charlie_test_a' [54 (real) element vector]

'charlie_test_b' [59 (real) element vector]

'charlie_test_c' [23 (real) element vector]

'jane_test_a' [24 (real) element vector]

'jane_test_b' [21 (real) element vector]

'mildred_test_a' [119 (real) element vector]

'mildred_test_b' [ 89 (real) element vector]

'mildred_test_c' [24 (real) element vector]

'pelagius_test_a' [314(real) element vector]

...

What I would like, is to concatenate the results vector of each subject so that I would get something (e.g. a struct concat_results) like

concat_results.charlie = [136 (real) element vector]

concat_results.jane = [45 (real) element vector]

concat_results.mildred = [232 (real) element vector]

Is there an intelligent/fast way to perform this grouped concatenation?

Thank you!

Answer 1

[Stephen23]

N = ["charlie_test_a";"charlie_test_b";"charlie_test_c";"jane_test_a";"jane_test_b";"mildred_test_a";"mildred_test_b";"mildred_test_c";"pelagius_test_a"]

N = 9×1 string array
    "charlie_test_a"
    "charlie_test_b"
    "charlie_test_c"
    "jane_test_a"
    "jane_test_b"
    "mildred_test_a"
    "mildred_test_b"
    "mildred_test_c"
    "pelagius_test_a"

D = {rand(1,54);rand(1,59);rand(1,23);rand(1,24);rand(1,21);rand(1,119);rand(1,89);rand(1,24);rand(1,314)};

The old-fashioned way:

[U,X,Y] = unique(extractBefore(N,"_"));
F = @(n)horzcat(D{Y==n});
C = arrayfun(F,1:max(Y),'uni',0);
S = cell2struct(C,U,2)

S = struct with fields:
     charlie: [1×136 double]
        jane: [1×45 double]
     mildred: [1×232 double]
    pelagius: [1×314 double]

Or using a table... not sure if this is much better:

T = table(N,D);
T.N = extractBefore(T.N,'_');
T = unstack(T, "D", "N", 'AggregationFunction',@(c){horzcat(c{:})});
T.charlie{1}

ans = 1×136

0.2737 0.3886 0.4304 0.4560 0.1101 0.8067 0.6387 0.3293 0.4217 0.7747 0.9228 0.6502 0.5434 0.1141 0.8645 0.9631 0.4089 0.3211 0.8731 0.3387 0.9358 0.8205 0.3495 0.0055 0.8159 0.0477 0.9314 0.7894 0.1157 0.4364

T.jane{1}

ans = 1×45

0.4863 0.7829 0.1432 0.3304 0.8526 0.4256 0.4426 0.4981 0.1038 0.4425 0.8619 0.7347 0.5812 0.5437 0.5414 0.8597 0.6437 0.7617 0.2114 0.0341 0.7536 0.7526 0.8522 0.2519 0.9126 0.9896 0.6591 0.0517 0.3433 0.6943

[Rafael Cordero]

Short, sweet and sexy.

Awesome!

Thank you!