Modified output_table to be created from cell array using cell2table() instead of array2table(). This extracts the contents of each cell, facilitating later column based filtering ops.

import/read_hdf5_as_table.m

@@ -58,9 +58,9 @@ end
 
 % Construct cell array with table data, where each row denotes a group, and
 % the columns are either datasets or attributes in the given group.
-tmp = {};
+tmp = cell(num_rows,num_cols); % {};
 idx = [];
-group_columns = {};
+group_columns =  {};
 for i = 1 : num_rows
     disp(['Begin construction of row:',INFO.Groups(i).Name])
 
@@ -86,7 +86,7 @@ for i = 1 : num_rows
             else % no data type conversion is performed
                 convert_func = @(x) x;
             end
-            
+
             tmp{i,j} = convert_func(h5readatt(filename,INFO.Groups(i).Name, Name)); 
             if i < 2
                 
@@ -115,5 +115,12 @@ end
 [~, idx_perm] = sort(idx,'ascend');
 
 varNames = [dataset_columns,group_columns];
-output_table = array2table(tmp(:,idx_perm), 'VariableNames', varNames(idx_perm));
+
+% Create output_table from cell array using cell2table. This as opposed to
+% array2table extract the content of its cells, which facilitates filtering
+% operations based on table columns/vars.
+
+%output_table = array2table(tmp(:,idx_perm), 'VariableNames', varNames(idx_perm));
+output_table = cell2table(tmp(:,idx_perm), 'VariableNames', varNames(idx_perm));
+