# 找出 q_table 中第二行, 列值最大的位置.[答案为 a2 或者 a3]
q_table
a1 a2 a3 a4
0 1 11 11 111
1 2 22 22 22
2 3 33 33 1
#------------------------
q_table = pd.DataFrame({'a1': [1, 2, 3], 'a2': [11, 22, 33], 'a3': [11, 22, 33], 'a4': [111, 22, 1] })
print("q_table \n", q_table)
print(" ------------ ")
state_action = q_table.loc[2, :]
print(" origin \n",state_action)
state_action = state_action.reindex(np.random.permutation(state_action.index))
print(" reindex \n", state_action)
print(" ------------ ")
action = state_action.idxmax()
print(" max value's index",action)
# 随机访问
action_index = np.random.choice(len(state_action))
action = state_action.index[action_index]