Horizontally concatenating table columns: combining a table column and a 2d array output by str2double() into one table

I'm reading in spreadsheet data from a file and storing that in the variable wab_table = readtable("WAB Summary Statistics Behavioral Data Language Scores.xlsx"); . The first column of this table contains subject ids, and I get the subject ids I need using the variable row_indices which contains the following.

8 14 10 9 2 4 7 3 6 11

There's also a variable i create called all_subjects which just has the typed out version of the subject ids at each of these row indices.

all_subjects = 
  10×1 string array
    "DE-008"
    "EX-014"
    "FL-010"
    "IS-009"
    "KI-002"
    "KX-004"
    "NS-007"
    "SH-003"
    "SN-006"
    "TO-011"

The following columns all have numerical data I'm interested in: [5;15;25;35;45]; . So I try to index into these columns at 10 selected row indices within the variable row_indices

avg_baselines_table_withnans = wab_table(row_indices,[5;15;25;35;45]); % unlabelled
avg_baselines_nonans = [];
for i=1:5
    col = str2double(avg_baselines_table_withnans.(i));
    avg_baselines_nonans = [avg_baselines_nonans, col];
end

avg_baselines_table_withnans is a subtable created using the indexing into wab_table as I wrote it above. Every column in avg_baselines_table_withnans is supposed to have numerical data but there are certain entries that have strings and study notes/things that can't be converted to a number.

I figured I needed to call str2double() on this subtable. But if I tried str2double(avg_baselines_table_withnans) this would output just one "Nan" for the whole table rather than converting each entry. I ultimately had to settle on the for loop I created above.

So at this point... I want to create a table that has the column for the subject ids first, and then it should have columns for each column in avg_baselines_nonans which should be the array created from passing each column from avg_baselines_table_withnans into str2double() .

Things work the way I think they should if I convert avg_baselines_nonans to a cell array and try to concatenate cell arrays by indexing into wab_table using {} operators and num2cell()

avg_baselines_cell_arr2 = [ wab_table{row_indices, 1}, num2cell(avg_baselines_nonans) ];
avg_baselines_table = array2table(avg_baselines_cell_arr2, 'VariableNames', {'Subj' 'avg_wabaq' 'avg_spont' 'avg_avc' 'avg_rep' 'avg_naming'})

This creates the table avg_baselines_table with columns listed as is in 'VariableNames'. Good.

I have a for loop that goes through a series of lines that use conditional indexing. Here is one of those lines. This works as I expect it to as long as I have concatenated cell arrays

subject_avg_baselines_row = avg_baselines_table(avg_baselines_table.Subj==all_subjects(i),:) % from the for loop
avg_baselines_table(avg_baselines_table.Subj==all_subjects(1),:)

What's confusing me is that these lines don't work as intended if, instead of using cell indexing/cell arrays, I just try to use normal arrays and normal indexing using () operators instead.

If I rewrite avg_baselines_cell_arr2 to the following...

avg_baselines_cell_arr2 = [wab_table(row_indices, 1), array2table(avg_baselines_nonans)];
avg_baselines_table = array2table(avg_baselines_cell_arr2, 'VariableNames', {'Subj' 'avg_wabaq' 'avg_spont' 'avg_avc' 'avg_rep' 'avg_naming'})

Then I get a table that actually seems to have two rows of headers?

array2table([wab_table(row_indices, 1), array2table(avg_baselines_nonans)], 'VariableNames', {'Subj' 'avg_wabaq' 'avg_spont' 'avg_avc' 'avg_rep' 'avg_naming'})

ans =

  10×6 table

      Subj            avg_wabaq                avg_spont                 avg_avc                  avg_rep                avg_naming      
      Var1      avg_baselines_nonans1    avg_baselines_nonans2    avg_baselines_nonans3    avg_baselines_nonans4    avg_baselines_nonans5

And if I try to proceed with this table, then my conditional statements such as avg_baselines_table(avg_baselines_table.Subj==all_subjects(1),:) don't work.

Instead they have to be changed to avg_baselines_table(avg_baselines_table.Subj.Var1==all_subjects(1),:) i.e. I have to use dot notation to refer to both column names or a dynamic selector like avg_baselines_table.Subj.(1)

So at this point I'm wondering

• if str2double() outputs a 2d array/matrix, how exactly do I merge/concatenate that array with the table column at wab_table(row_indices, 1) ? Do I have to use cell indexing, or is there a way to use normal indexing and for things to still work?
• I'm calling array2table twice because if you use the concatenor operators [] then everything within those operators needs to be a table otherwise the concatenation fails/gives an error. So I'm calling array2table around the concatenated table columns of interest and I'm calling array2table within the concatenation to convert the array produced by str2double into a table. It seems like this is what's causing me to have this problem. Is there a way I can concatenate the column containing subject ids with the output of str2double so that I don't have to worry about having two variable headers?