06_索引和切片

索引和切片

• 数组切片是原始数组的视图，不是复制品，如果要得到切片的副本，应该使用copy()方法
• 在多维数组中索引，如果省略了后面的索引，则返回对象会是一个维度低一点的ndarray（它含有高一级维度上的所有数据）

1、切片

切片的基本语法 ndarray[i : j : k]，其中i为起始下标，j为结束下标（不包括j），k为步长（默认为1）

import numpy as np

arr = np.arange(10)
arr

>>> array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])

---

arr[1:9:2]

>>> array([1, 3, 5, 7])

---

arr[4:-3]

>>> array([4, 5, 6])

---

arr[9:4:-2]

>>> array([9, 7, 5])

---

a = arr[::2]
a

>>> array([0, 2, 4, 6, 8])

---

arr[:6] = 0

>>> array([0, 0, 0, 0, 0, 0, 6, 7, 8, 9])

---

arr[6:]

>>> array([6, 7, 8, 9])

---

arr[:] = 1
arr

>>> array([1, 1, 1, 1, 1, 1, 1, 1, 1, 1])

---

多维切片

ndarr = np.arange(16).reshape((4, 4))
ndarr

>>> array([[ 0,  1,  2,  3],
           [ 4,  5,  6,  7],
           [ 8,  9, 10, 11],
           [12, 13, 14, 15]])

---

ndarr[:2] # 行，等效于arr2[:2, :]

>>> array([[0, 1, 2, 3],
           [4, 5, 6, 7]])

---

ndarr[:2, :2] # 前两行的前两列

>>> array([[0, 1],
           [4, 5]])

---

ndarr[:, -2] 

>>> array([ 2, 6, 10, 14])

---

ndarr[::2, -2] # 步长为2的行的倒数第二列

>>> array([ 2, 10])

---

2、索引

arr = np.arange(10)
ndarr = np.arange(16).reshape((4, 4))
arr
ndarr

>>> array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
>>> array([[ 0,  1,  2,  3],
           [ 4,  5,  6,  7],
           [ 8,  9, 10, 11],
           [12, 13, 14, 15]])

---

arr[1]

>>> 1

---

ndarr[1,1] # 等效于ndarr[1][1]，但是后者速度更慢

>>> 5

---

ndarr[1][1]

>>> 5

---

ndarr[2, [1, 2]] # 第2行的第1、2列元素

>>> array([ 9, 10])

---

ndarr[:3, [0,1]] # 前3行的第0、1列元素

>>> array([[0, 1],
           [4, 5],
           [8, 9]])

---

布尔索引

cities = np.array(['bj', 'cd', 'sh', 'gz', 'cd'])
cities

>>> array(['bj', 'cd', 'sh', 'gz', 'cd'], dtype='<U2')

---

data = np.arange(20).reshape(5, 4)
data

>>> array([[ 0,  1,  2,  3],
           [ 4,  5,  6,  7],
           [ 8,  9, 10, 11],
           [12, 13, 14, 15],
           [16, 17, 18, 19]])

---

cities == 'cd'

>>> array([False,  True, False, False,  True])

---

data[cities == 'cd']

>>> array([[ 4,  5,  6,  7],
           [16, 17, 18, 19]])

---

data[cities == 'cd', :1]

>>> array([[ 4],
           [16]])

---

cities != 'cd'

>>> array([ True, False,  True,  True, False])

---

data[cities != 'cd']  # 等效于 data[~(cities == 'cd')] 

>>> array([[ 0,  1,  2,  3],
           [ 8,  9, 10, 11],
           [12, 13, 14, 15]])

---

data[cities != 'cd', :1] # 等效于 data[~(cities == 'cd'), :1]

>>> array([[ 0],
           [ 8],
           [12]])

---

花式索引

指的是利用整数数组进行索引

arr3 = np.arange(36).reshape(6,6) 
arr3

>>> array([[ 0,  1,  2,  3,  4,  5],
           [ 6,  7,  8,  9, 10, 11],
           [12, 13, 14, 15, 16, 17],
           [18, 19, 20, 21, 22, 23],
           [24, 25, 26, 27, 28, 29],
           [30, 31, 32, 33, 34, 35]])

---

# 取出第3，2, 5行
arr3[[3, 4, 5]]

>>> array([[18, 19, 20, 21, 22, 23],
           [24, 25, 26, 27, 28, 29],
           [30, 31, 32, 33, 34, 35]])

---

# 使用负数索引将会从末尾开始选取行
arr3[[-4, -2 , -1]]

>>> array([[12, 13, 14, 15, 16, 17],
           [24, 25, 26, 27, 28, 29],
           [30, 31, 32, 33, 34, 35]])

---

arr3[[1, 2, 3], [1, 2, 3]] # 取出下标为(1, 1), (2, 2), (3, 3)的元素 

>>> array([ 7, 14, 21])

---

arr3[[1, 2, 3]][:, [1, 2, 3]]  # 同时在行和列上进行花式索引

>>> array([[ 7,  8,  9],
           [13, 14, 15],
           [19, 20, 21]])

---