numpy数组的索引和切片

提示：扫一扫查出行
【扫一扫了解最新限行尾号】
复制提示

基本切片操作

>>> import numpy as np >>> arr=np.arange(10) >>> arr array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]) >>> arr[5] 5 >>> arr[5:8] array([5, 6, 7])

切片赋值操作

1.切片赋一个值对应原来数组中的值也会变

>>> arr[5:8]=12 >>> arr array([ 0, 1, 2, 3, 4, 12, 12, 12, 8, 9])

>>> import numpy as np >>> arr=np.arange(10) >>> arr_slice=arr[5:8] >>> arr_slice[0]=-1 >>> arr_slice array([-1, 6, 7]) >>> arr array([ 0, 1, 2, 3, 4, -1, 6, 7, 8, 9])

2.给数组中所有元素赋值

>>> arr[:]=-1 >>> arr array([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1])

3.如果想使用复制的方法,使用copy方法

>>> arr_copy=arr[:].copy() >>> arr_copy array([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) >>> arr_copy[:]=0 >>> arr_copy array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0]) >>> arr array([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1])

高阶数组索引

>>> import numpy as np >>> arr2d=np.array([[1,2,3],[4,5,6],[7,8,9]]) >>> arr2d[2] array([7, 8, 9]) >>> arr2d[0][2] 3 >>> arr2d[0,2] 3

>>> import numpy as np >>> arr2d=np.array([[1,2,3],[4,5,6],[7,8,9]]) >>> arr2d[2] array([7, 8, 9]) >>> arr2d[0][2] 3 >>> arr2d[0,2] 3 >>> arr3d=np.array([[[1,2,3],[4,5,6]],[[7,8,9],[10,11,12]]]) >>> arr3d array([[[ 1, 2, 3], [ 4, 5, 6]], [[ 7, 8, 9], [10, 11, 12]]]) >>> arr3d[0] array([[1, 2, 3], [4, 5, 6]]) >>> old_values=arr3d[0].copy() >>> arr3d[0]=42 >>> arr3d array([[[42, 42, 42], [42, 42, 42]], [[ 7, 8, 9], [10, 11, 12]]]) >>> arr3d[1,0] array([7, 8, 9]) >>> x=arr3d[1] >>> x array([[ 7, 8, 9], [10, 11, 12]]) >>> x[0] array([7, 8, 9])

高维数组切片

>>> arr2d[:2] array([[1, 2, 3], [4, 5, 6]]) >>> arr2d[:2,1:] array([[2, 3], [5, 6]]) >>> arr2d[1,:2] array([4, 5]) >>> arr2d[:2,2] array([3, 6]) >>> arr2d[:,:1] array([[1], [4], [7]])

布尔型索引

1.假设我们有一个用于存储数据的数组以及一个存储姓名的数组（含有重复项）。在这里，我将使用numpy.random中的randn函数生成一些正态分布的随机数据：

>>> import numpy as np >>> names = np.array(['Bob', 'Joe', 'Will', 'Bob', 'Will', 'Joe', 'Joe']) >>> data=np.random.randn(7,4)#7行4列正太分布随机数组 >>> names array(['Bob', 'Joe', 'Will', 'Bob', 'Will', 'Joe', 'Joe'], dtype='<U4') >>> data array([[ 0.24724057, 2.86939948, -0.82061782, -0.65745818], [-0.98602372, -0.69305692, -1.44431904, -0.85490816], [-0.73613349, 0.12700976, -1.00588979, 1.10646269], [ 1.59110894, 1.68597758, 0.39414277, 2.02308399], [-1.05607115, -0.50354292, -0.65820553, -0.77610316], [ 1.72237936, -0.07726577, 1.63462647, -0.41943148], [ 0.66744687, -1.01756773, -0.59254343, 0.19080575]])

2.假设每个名字都对应data数组中的一行，而我们想要选出对应于名字"Bob"的所有行。跟算术运算一样，数组的比较运算（如==）也是矢量化的。因此，对names和字符串"Bob"的比较运算将会产生一个布尔型数组：

>>> names=='Bob' array([ True, False, False, True, False, False, False])

3.布尔数组可以用于数组的索引

获取等于'Bob'的行

>>> data[names=='Bob'] array([[ 0.24724057, 2.86939948, -0.82061782, -0.65745818], [ 1.59110894, 1.68597758, 0.39414277, 2.02308399]])

获取不同于'Bob'的行

>>> data[names!='Bob'] array([[-0.98602372, -0.69305692, -1.44431904, -0.85490816], [-0.73613349, 0.12700976, -1.00588979, 1.10646269], [-1.05607115, -0.50354292, -0.65820553, -0.77610316], [ 1.72237936, -0.07726577, 1.63462647, -0.41943148], [ 0.66744687, -1.01756773, -0.59254343, 0.19080575]])

4.对布尔索引进行列索引

>>> data[names=='Bob',2:] array([[-0.82061782, -0.65745818], [ 0.39414277, 2.02308399]]) >>> data[names=='Bob',3] array([-0.65745818, 2.02308399])

5.反转条件符

>>> cond=names=='Will' >>> cond array([False, False, True, False, True, False, False]) >>> data[~cond] array([[ 0.24724057, 2.86939948, -0.82061782, -0.65745818], [-0.98602372, -0.69305692, -1.44431904, -0.85490816], [ 1.59110894, 1.68597758, 0.39414277, 2.02308399], [ 1.72237936, -0.07726577, 1.63462647, -0.41943148], [ 0.66744687, -1.01756773, -0.59254343, 0.19080575]])

6.布尔条件的运算

除此之外，连接符还有|、&之类

>>> mask=(names=='Bob')|(names=='Will') >>> mask array([ True, False, True, True, True, False, False]) >>> data[mask] array([[ 0.24724057, 2.86939948, -0.82061782, -0.65745818], [-0.73613349, 0.12700976, -1.00588979, 1.10646269], [ 1.59110894, 1.68597758, 0.39414277, 2.02308399], [-1.05607115, -0.50354292, -0.65820553, -0.77610316]])

7.条件选取

普通条件选取

>>> data[data<0]=0 >>> data array([[0.24724057, 2.86939948, 0. , 0. ], [0. , 0. , 0. , 0. ], [0. , 0.12700976, 0. , 1.10646269], [1.59110894, 1.68597758, 0.39414277, 2.02308399], [0. , 0. , 0. , 0. ], [1.72237936, 0. , 1.63462647, 0. ], [0.66744687, 0. , 0. , 0.19080575]])

布尔条件选取

>>> import numpy as np >>> names = np.array(['Bob', 'Joe', 'Will', 'Bob', 'Will', 'Joe', 'Joe']) >>> data=np.random.randn(7,4)#7行4列正太分布随机数组 >>> data array([[-1.24077681, -0.48320904, 1.22145611, 0.00666619], [-0.65078721, -0.03482355, 1.74232625, 0.2979584 ], [-1.51669752, 2.04245014, 0.09453898, -0.85531867], [-1.51334497, 0.36947066, -0.87016919, 1.35107873], [-1.11285867, -2.20906849, 0.38269412, 1.85375798], [ 0.95132554, -1.54193589, 1.98741745, -0.60608077], [ 0.78902133, 1.41593836, 0.09430052, -0.25057659]]) >>> data[names!='Joe']=7 >>> data array([[ 7. , 7. , 7. , 7. ], [-0.65078721, -0.03482355, 1.74232625, 0.2979584 ], [ 7. , 7. , 7. , 7. ], [ 7. , 7. , 7. , 7. ], [ 7. , 7. , 7. , 7. ], [ 0.95132554, -1.54193589, 1.98741745, -0.60608077], [ 0.78902133, 1.41593836, 0.09430052, -0.25057659]]) >>>

花式索引

1.传入单个索引数组

>>> import numpy as np >>> arr=np.empty((8,4))#创建8行4列内容为随机值的数组 >>> arr array([[2.65577744e-260, 7.70858946e+218, 6.01334668e-154, 4.47593816e-091], [7.01413727e-009, 2.96905203e+222, 2.11672643e+214, 4.56532297e-085], [4.78409596e+180, 2.44001263e-152, 2.45981714e-154, 6.83528875e+212], [6.14829725e-071, 1.05161522e-153, 1.05135742e-153, 2.43902457e-154], [4.83245960e+276, 6.03103052e-154, 7.06652000e-096, 2.65862875e-260], [1.76380220e+241, 2.30576063e-310, 9.80013217e+040, 1.55850644e-312], [1.33360318e+241, 4.09842267e-310, 2.48721655e-075, 1.04922745e-312], [1.91217285e-309, 1.18182126e-125, 6.57144273e-299, 5.54240979e-302]]) >>> for i in range(8): arr[i]=i>>> arr array([[0., 0., 0., 0.], [1., 1., 1., 1.], [2., 2., 2., 2.], [3., 3., 3., 3.], [4., 4., 4., 4.], [5., 5., 5., 5.], [6., 6., 6., 6.], [7., 7., 7., 7.]]) >>> arr[[4,3,0,6]]#选特定的索引下标,选取第4，3，0，6行 array([[4., 4., 4., 4.], [3., 3., 3., 3.], [0., 0., 0., 0.], [6., 6., 6., 6.]]) >>> arr[[-3,-5,-7]]#选择特定的索引下标，选取第-3,-5,-7列 array([[5., 5., 5., 5.], [3., 3., 3., 3.], [1., 1., 1., 1.]])

2.传入多个索引数组

>>> arr=np.arange(32).reshape((8,4)) >>> arr 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]]) >>> arr[[1,5,7,2],[0,3,1,2]]#选取(1,0),(5,3),(7,1),(2,2)对应元素 array([ 4, 23, 29, 10])

>>> arr[[1,5,7,2]][:,[0,3,1,2]]#先选取第1,5,7,2行，再将每行按照0,3,1,2这个顺序交换 array([[ 4, 7, 5, 6], [20, 23, 21, 22], [28, 31, 29, 30], [ 8, 11, 9, 10]])