How do I sample few numbers with different distributions?

Let say I have my own custom distribution numpy vector called p.

Then p satisfy the next: np.ndim(p) == 1 & np.sum(p) == 1 & np.all(p >= 0)

With that vector I can easily sample a number in 0 - p.shape with np.random.choice(np.arange(len(p)), p=p)

In a case I have many such ps, I have a matrix (with dim 2) P that satisfies:

np.sum(P[:,i]) == 1   # for all i in P.shape[1]
np.all(P >= 0)

Then I wish to sample P.shape[1] numbers in the range 0 to P.shape[0] with probability P.

For example the next code:

P = np.array([[0.2, 0.3],
              [0.5, 0.7],
              [0.3, 0]])
x = np.random.choice(np.arange(P.shape[0], P[:,0]))
y = np.random.choice(np.arange(P.shape[0], P[:,1]))

will produce my will (x=0 in 0.2, x=1 in 0.5 and x=2 in 0.3 and y=0 in 0.3, y=1 in 0.7).

In my case P has many columns and I wish to sample all in one shot.

Of course I can do it in a for loop, for example:

random_values = np.empty(P.shape[1])
arange_arr = np.arange(P.shape[0])
for i in range(P.shape[1]):
    random_values[i] = np.random.choice(arnge_arr, p=P[:,i])

Trying to find some nupmy-scipy elegant way to do it.

Thanks