Here's one way to do it:
% construct the initial table:
Items = ["Item1";"Item2";"Item3";"Item4";"Item5";"Item6";"Item7";"Item8"];
Family = ["Family 1";"Family 2";"Family 4";"Family 2";"Family 3";"Family 1";"Family 1";"Family 3"];
t = table(Items,Family)
% group t.Family:
[idx,group_ids] = findgroups(t.Family);
n_groups = numel(group_ids); % number of Family groups
% split t.Items into groups according to Family:
grouped_items = splitapply(@(x){x},t.Items,idx);
n_items = cellfun(@numel,grouped_items); % number of Items in each Family group
% construct 2d string array of Item* Strings, where each column
% is the Items in a Family group, filled out with <missing> String(s):
vals = repmat(string(missing()),max(n_items),n_groups);
for ii = 1:n_groups
vals(1:n_items(ii),ii) = grouped_items{ii};
end
% convert 2d String array into a cell array,
% to be used as an argument list in table():
vals = num2cell(vals,1);
% create the new table:
new_t = table(vals{:},'VariableNames',group_ids)
