Best way to write bootstrapping in python

I was programming a method that applied statistical bootstrapping over a sample in python, and I have come with two solutions, one which is fully vectorized, and other that uses list comprehension.

import seaborn as sns
import numpy as np
import matplotlib.pyplot as plt

sample = np.array([80,75,91,97,88,77,94])

bs_sample_1 = np.random.choice(sample,size = (10000,7)).mean(axis = 1)

bs_sample_2 = np.array([np.random.choice(sample,size = 7).mean() for i in range(10000)])

plt.figure()
sns.distplot(sample)

plt.figure()
sns.distplot(bs_sample_1)

plt.figure()
sns.distplot(bs_sample_2)

I don't have much knowledge about RNG, but I not sure if the two operations are equally valid to generate bootstrap samples.