[FIXED] how to avoid append many times?

Issue

I have a piece of code which involves many append command,

 items = [a1,a2,a3,a4,a5,a6,a7]
 A_all = []
 for t in range(len(items)):
     A_BS = []
     for i in range(iter):
         idx_BS = np.random.choice(np.shape(items[0])[1], boots)
         corr_atra = get_corr((items[t][:,idx_BS]).T)
         A_BS.append(np.mean(corr_atra))
 A_all.append(np.mean(A_BS)) 

what the code do is that for each data set a_i bootstrap among columns and compute the correlation and then get the average there. Is there a way to avoid some of these appends?

Solution

You could pack it into list comprehensions:

s = np.shape(items[0])[1]
A_all = [
    [
        np.mean(get_corr(item[:, np.random.choice(s, boots)].T))
        for _ in range(iter)
    ]
    for item in items
]

But: Whatever iter is should have another name because iter is a built-in function.

You don't use i - is that intensional?

Answered By - Timus

Answer Checked By - Pedro (PHPFixing Volunteer)