Splitting a table to 2

Hello all, I have a table named A. I want to split it into 2 sub-tables: Table 1 has 60% of the rows in A, chosen randomly. Table 2 has all the other rows that are not in table 1. Any elegant idea on how to do this?

Answer 1

[KL]

dummy = 10*rand(10,4);
T = array2table(dummy)
ranRows = randperm(length(dummy));
T1 = T(ranRows(1:(0.6*length(dummy))),:)
T2 = T(ranRows((0.6*length(dummy))+1:end),:)

[Guillaume]

The proper function for getting the height of a table is height, so:

ranrows = randperm(height(T));
T1 = T(ranRows(1:0.6*height(T)), :);
T2 = T(ranRows(0.6*height(T)+1:end), :);

To avoid a warning, I would also wrap the 0.6*height(T) calculation inside a ceil, a floor or a round.

[KL]

Thanks Guillaume!

Answer 2

[Adrian]

Note that this will only play nicely if you have a reasonably large array.

selection = randi([1 10], size(A,1),1)
A_sixty = A(selection>4);
A_forty = A(selection<=4);

Answer 3

[Sean de Wolski]

If you're looking to split your data into a training and testing set, then it might pay to look at cvpartition as well, it provides more options for splitting tables.