different behaviour when concatenating table colums

The two lines behave differently:

tmp = [all_table.('SpO2_FiO2');myTable.('SpO2_FiO2')]; % this works
tmp = [all_table(:,52);myTable(:,52)]; % this gives error 'Cannot concatenate the table variable 'SpO2_FiO2' because it is a cell in one table and a non-cell in another.'

52 is the column index to the column named ''SpO2_FiO2". The type of this column is double in both tables, all_table and my_Table.

Answer 1

[Voss]

If both tables' SpO2_FiO2 columns are actually doubles, then both ways of concatenating do the same thing:

SpO2_FiO2 = (1:3).';

all_table = table(SpO2_FiO2)

all_table = 3×1 table
    SpO2_FiO2
    _________

        1    
        2    
        3    

myTable = table(SpO2_FiO2)

myTable = 3×1 table
    SpO2_FiO2
    _________

        1    
        2    
        3    

tmp = [all_table.('SpO2_FiO2');myTable.('SpO2_FiO2')]

tmp = 6×1

1 2 3 1 2 3

tmp = [all_table(:,1);myTable(:,1)]

tmp = 6×1 table
    SpO2_FiO2
    _________

        1    
        2    
        3    
        1    
        2    
        3    

If instead one table's SpO2_FiO2 column is a cell array, then you get behavior consistent with what you observe:

SpO2_FiO2 = (1:3).';

all_table = table(SpO2_FiO2)

all_table = 3×1 table
    SpO2_FiO2
    _________

        1    
        2    
        3    

SpO2_FiO2 = num2cell(SpO2_FiO2);
myTable = table(SpO2_FiO2)

myTable = 3×1 table
    SpO2_FiO2
    _________

      {[1]}  
      {[2]}  
      {[3]}  

tmp = [all_table.('SpO2_FiO2');myTable.('SpO2_FiO2')]

tmp = 4×1 cell array
    {3×1 double}
    {[       1]}
    {[       2]}
    {[       3]}

tmp = [all_table(:,1);myTable(:,1)]

Cannot concatenate the table variable 'SpO2_FiO2' because it is a cell in one table and a non-cell in another.

In the second case, the line:

tmp = [all_table.('SpO2_FiO2');myTable.('SpO2_FiO2')];

does not generate an error, but it doesn't necessarily 'work' because it doesn't necessarily do what you want. It gives a 4-by-1 cell array where the first element is a 3-by-1 vector of doubles and the 2nd, 3rd, and 4th elements are all scalar doubles, as opposed to a 6-by-1 vector of doubles, which is what you get either way you do the concatenation when both columns are actually doubles.