A for loop extracting every new variable into column vectors?

Hi! I have the following stacked table, and I need to be able to extract/refer to every new element that appears at ChElem_Indicator (column 8) with the correspondent value that appears at ChElem (column 9). The ChElem_Indicator contains repeated data, with the elements (for example "Si") appearing hundreds of times.

Every time a new ChElem_Indicator variable appears, I'd like to create a column vector (with the name of that variable) that extracts the corresponding value from ChElem column. Scanning the entire table, my goal is to end up with, for example, a column vector "Si" with all values which are related to "Si", and to have this for all the different variables that appear in ChElem_Indicator column (not only "Si"). I think a for loop is what I need, but I don't know how to incorporate the part "for every new variable".

Any ideas/suggestions? Thanks in advance!

[Stephen23]

'I think a for loop is what I need, but I don't know how to incorporate the part "for every new variable".'

Do NOT do this! Magically accessing variable names in a loop is how beginners force themselves into writing slow, complex, buggy code:

https://www.mathworks.com/matlabcentral/answers/304528-tutorial-why-variables-should-not-be-named-dynamically-eval

'my goal is to end up with, for example, a column vector "Si" with all values which are related to "Si", and to have this for all the different variables that appear in ChElem_Indicator column (not only "Si").'

'Any ideas/suggestions?'

Use one of the table class methods to group the rows based on that column, such as

https://www.mathworks.com/help/matlab/ref/unstack.html

or

https://www.mathworks.com/help/matlab/ref/splitapply.html

[Ana Castanheiro]

Thanks for sharing your advices and suggestions, these are especially important for a Matlab beginner as myself.

Answer 1

[Guillaume]

[groupid, groupname] = findgroups(ClustSEMdata.ChElem_Indicator);
elements = splitapply(@(v) {v}, ClustSEMdata.ChElem, groupid);

result = cell2table(elements.', 'VariableNames', groupname)

It sounds like you want to create individual variables for each element, which would be a very bad idea. Putting them in a table as above, a cell array, a map, or a structure would be a lot better.

[Ana Castanheiro]

Many thanks for your reply Guillaume. This piece of code works nicely but only partially for what I was asking about. Based on your and Stephen's comments, I now understand that creating individual variables for each element is not a good idea. I had created the table in the image using the stack function, but probably this is not necessary at all, and I should instead work with the original table (where every column is already a different chemical element).

What I need is to be able to do some clustering analysis, for which I need to start with the code below. So, perhaps the best is that I use:

treedata = [ originaldata.Al, originaldata.Si, originaldata.Fe];

 %%desired treedata = [Al, Si, Fe]; 
dist = pdist(treedata,'euclidean');
linkmethod = 'ward';
tree = linkage(dist,linkmethod);
figure()
dendrogram(tree, 0, 'orientation', 'right')
title(sprintf('%s%s %d', det1{1:2},det1{3}))