参考
https://docs.scipy.org/doc/numpy-dev/user/quickstart.html
基础
NumPy的主要对象是齐次多维数组。它是一个元素表（通常是数字），所有相同的类型，由正整数的元组索引。在NumPy维度被称为轴（axis）。轴的数量是等级（rank）。
例如，三维空间中一个点的坐标[1,2,1]是一个等级为1的数组，因为它具有一个坐标轴。该轴的长度为3.在下面的示例中，该数组具有等级2（它是二维的）。第一维（轴）的长度为2，第二维的长度为3。
[[ 1. ， 0. ， 0 ]，
[ 0. ， 1. ， 2. ]]
NumPy的数组类叫做ndarray，别名为array，有几个重要的属性
ndarray.ndim ：维度
ndarray.shape ：尺寸，如n行m列（n，m）
ndarray.size：元素总数
ndarray.dtype：一个描述数组中元素类型的对象。可以使用标准的Python类型创建或指定dtype。另外NumPy提供它自己的类型。numpy.int32，numpy.int16和numpy.float64是一些例子。
ndarray.itemsize：数组中每个元素的字节大小
ndarray.data：包含数组的实际元素的缓冲区.
>>> import numpy as np
>>> a = np.arange(15).reshape(3, 5)
>>> a
array([[ 0, 1, 2, 3, 4],
[ 5, 6, 7, 8, 9],
[10, 11, 12, 13, 14]])
>>> a.shape
(3, 5)
>>> a.ndim
2
>>> a.dtype.name
'int64'
>>> a.itemsize
8
>>> a.size
15
>>> type(a)
<type 'numpy.ndarray'>
>>> b = np.array([6, 7, 8])
>>> b
array([6, 7, 8])
>>> type(b)
<type 'numpy.ndarray'>
创建数组
>>> import numpy as np
>>> a = np.array([2,3,4]) #内部小括号，中括号都可以
>>> a
array([2, 3, 4])
>>> a.dtype
dtype('int64')
>>> b = np.array([1.2, 3.5, 5.1])
>>> b.dtype
dtype('float64')
容易犯的错
>>> a = np.array(1,2,3,4) # WRONG
>>> a = np.array([1,2,3,4]) # RIGHT
二位数组创建
>>> b = np.array([(1.5,2,3), (4,5,6)])
>>> b
array([[ 1.5, 2. , 3. ],
[ 4. , 5. , 6. ]])
创建时可以指定元素类型
>>> c = np.array( [ [1,2], [3,4] ], dtype=complex )
>>> c
array([[ 1.+0.j, 2.+0.j],
[ 3.+0.j, 4.+0.j]])
快捷创建函数
>>> np.zeros( (3,4) )
array([[ 0., 0., 0., 0.],
[ 0., 0., 0., 0.],
[ 0., 0., 0., 0.]])
>>> np.ones( (2,3,4), dtype=np.int16 ) # dtype can also be specified
array([[[ 1, 1, 1, 1],
[ 1, 1, 1, 1],
[ 1, 1, 1, 1]],
[[ 1, 1, 1, 1],
[ 1, 1, 1, 1],
[ 1, 1, 1, 1]]], dtype=int16)
>>> np.empty( (2,3) ) # uninitialized, output may vary
array([[ 3.73603959e-262, 6.02658058e-154, 6.55490914e-260],
[ 5.30498948e-313, 3.14673309e-307, 1.00000000e+000]])
类比range函数可以使用以下方式创建，返回数组而不是list
>>> np.arange( 10, 30, 5 )
array([10, 15, 20, 25])
>>> np.arange( 0, 2, 0.3 ) # it accepts float arguments
array([ 0. , 0.3, 0.6, 0.9, 1.2, 1.5, 1.8])
上述方法不能预订元素个数，可以通过以下方法创建固定元素个数的数组
>>> from numpy import pi
>>> np.linspace( 0, 2, 9 ) # 9 numbers from 0 to 2
array([ 0. , 0.25, 0.5 , 0.75, 1. , 1.25, 1.5 , 1.75, 2. ])
>>> x = np.linspace( 0, 2*pi, 100 ) # useful to evaluate function at lots of points
>>> f = np.sin(x)
创建数组的方法
array，zeros，zeros_like,ones,ones_like,empty,empty_like,arange,linspace,numpy.random.rand,numpy.random.randn,fromfunction,fromfile
打印数组
print函数打印，具体打印方式看代码结果
>>> a = np.arange(6) # 1d array
>>> print(a)
[0 1 2 3 4 5]
>>>
>>> b = np.arange(12).reshape(4,3) # 2d array
>>> print(b)
[[ 0 1 2]
[ 3 4 5]
[ 6 7 8]
[ 9 10 11]]
>>>
>>> c = np.arange(24).reshape(2,3,4) # 3d array
>>> print(c)
[[[ 0 1 2 3]
[ 4 5 6 7]
[ 8 9 10 11]]
[[12 13 14 15]
[16 17 18 19]
[20 21 22 23]]]
如果打印过多，会自动省略一部分
>>> print(np.arange(10000))
[ 0 1 2 ..., 9997 9998 9999]
>>>
>>> print(np.arange(10000).reshape(100,100))
[[ 0 1 2 ..., 97 98 99]
[ 100 101 102 ..., 197 198 199]
[ 200 201 202 ..., 297 298 299]
...,
[9700 9701 9702 ..., 9797 9798 9799]
[9800 9801 9802 ..., 9897 9898 9899]
[9900 9901 9902 ..., 9997 9998 9999]]
如果不想省略，改变打印参数set_printoptions
>>> np.set_printoptions(threshold='nan')
基本运算
>>> a = np.array( [20,30,40,50] )
>>> b = np.arange( 4 )
>>> b
array([0, 1, 2, 3])
>>> c = a-b
>>> c
array([20, 29, 38, 47])
>>> b**2
array([0, 1, 4, 9])
>>> 10*np.sin(a)
array([ 9.12945251, -9.88031624, 7.4511316 , -2.62374854])
>>> a<35
array([ True, True, False, False], dtype=bool)
如果想使用类似矩阵乘法运算，可使用dot函数
>>> A = np.array( [[1,1],
... [0,1]] )
>>> B = np.array( [[2,0],
... [3,4]] )
>>> A*B # elementwise product
array([[2, 0],
[0, 4]])
>>> A.dot(B) # matrix product
array([[5, 4],
[3, 4]])
>>> np.dot(A, B) # another matrix product
array([[5, 4],
[3, 4]])
也可以使用+= ，*=等运算
>>> a = np.ones((2,3), dtype=int)
>>> b = np.random.random(2,3)
>>> a *= 3
>>> a
array([[3, 3, 3],
[3, 3, 3]])
>>> b += a
>>> b
array([[ 3.417022 , 3.72032449, 3.00011437],
[ 3.30233257, 3.14675589, 3.09233859]])
>>> a += b # b is not automatically converted to integer type
Traceback (most recent call last):
...
TypeError: Cannot cast ufunc add output from dtype('float64') to dtype('int64') with casting rule 'same_kind'
不同元素类型的数组运算，类型自动上升转换
>>> a = np.ones(3, dtype=np.int32)
>>> b = np.linspace(0,pi,3)
>>> b.dtype.name
'float64'
>>> c = a+b
>>> c
array([ 1. , 2.57079633, 4.14159265])
>>> c.dtype.name
'float64'
>>> d = np.exp(c*1j)
>>> d
array([ 0.54030231+0.84147098j, -0.84147098+0.54030231j,
-0.54030231-0.84147098j])
>>> d.dtype.name
'complex128'
很多一元运算可以使用函数计算
>>> a = np.random.random(2,3)
>>> a
array([[ 0.18626021, 0.34556073, 0.39676747],
[ 0.53881673, 0.41919451, 0.6852195 ]])
>>> a.sum()
2.5718191614547998
>>> a.min()
0.1862602113776709
>>> a.max()
0.6852195003967595
可以指定axis来指定某一维度运算
>>> b = np.arange(12).reshape(3,4)
>>> b
array([[ 0, 1, 2, 3],
[ 4, 5, 6, 7],
[ 8, 9, 10, 11]])
>>>
>>> b.sum(axis=0) # sum of each column
array([12, 15, 18, 21])
>>>
>>> b.min(axis=1) # min of each row
array([0, 4, 8])
>>>
>>> b.cumsum(axis=1) # cumulative sum along each row
array([[ 0, 1, 3, 6],
[ 4, 9, 15, 22],
[ 8, 17, 27, 38]])
universal function
也可以使用数学运算，在numpy中叫做“universal functions”（u func）
>>> B = np.arange(3)
>>> B
array([0, 1, 2])
>>> np.exp(B)
array([ 1. , 2.71828183, 7.3890561 ])
>>> np.sqrt(B)
array([ 0. , 1. , 1.41421356])
>>> C = np.array([2., -1., 4.])
>>> np.add(B, C)
array([ 2., 0., 6.])
其他的看参考原文
索引，分片，遍历
一维数组与list相似
>>> a = np.arange(10)**3
>>> a
array([ 0, 1, 8, 27, 64, 125, 216, 343, 512, 729])
>>> a[2]
8
>>> a[2:5]
array([ 8, 27, 64])
>>> a[:6:2] = -1000 # equivalent to a[0:6:2] = -1000; from start to position 6, exclusive, set every 2nd element to -1000
>>> a
array([-1000, 1, -1000, 27, -1000, 125, 216, 343, 512, 729])
>>> a[ : :-1] # reversed a
array([ 729, 512, 343, 216, 125, -1000, 27, -1000, 1, -1000])
>>> for i in a:
... print(i**(1/3.))
...
nan
1.0
nan
3.0
nan
5.0
6.0
7.0
8.0
9.0
多维度数组每个轴可以有一个索引。这些索引用逗号分隔
>>> def f(x,y):
... return 10*x+y
...
>>> b = np.fromfunction(f,(5,4),dtype=int)
>>> b
array([[ 0, 1, 2, 3],
[10, 11, 12, 13],
[20, 21, 22, 23],
[30, 31, 32, 33],
[40, 41, 42, 43]])
>>> b[2,3]
23
>>> b[0:5, 1] # each row in the second column of b
array([ 1, 11, 21, 31, 41])
>>> b[ : ,1] # equivalent to the previous example
array([ 1, 11, 21, 31, 41])
>>> b[1:3, : ] # each column in the second and third row of b
array([[10, 11, 12, 13],
[20, 21, 22, 23]])
当提供的索引数量少于轴数时，缺少的索引被认为是完整的切片：
>>> b[-1] # the last row. Equivalent to b[-1,:]
array([40, 41, 42, 43])
维度过多时可以使用“…”来表示。
例如一个五维数组
- x[1,2,…] 等同于 x[1,2,:,:,:],
- x[…,3] 等同 x[:,:,:,:,3]
- x[4,…,5,:] 等同 x[4,:,:,5,:].
>>> c = np.array( [[[ 0, 1, 2], # a 3D array (two stacked 2D arrays)
... [ 10, 12, 13]],
... [[100,101,102],
... [110,112,113]]])
>>> c.shape
(2, 2, 3)
>>> c[1,...] # same as c[1,:,:] or c[1]
array([[100, 101, 102],
[110, 112, 113]])
>>> c[...,2] # same as c[:,:,2]
array([[ 2, 13],
[102, 113]])
遍历多维度数组遵从第一维度
>>> for row in b:
... print(row)
...
[0 1 2 3]
[10 11 12 13]
[20 21 22 23]
[30 31 32 33]
[40 41 42 43]
如果想一个一个元素遍历，可以使用flat（扁平化）属性
>>> for element in b.flat:
... print(element)
...
0
1
2
3
10
11
12
13
20
21
22
23
30
31
32
33
40
41
42
43
##数组形状操作
改变形状
>>> a = np.floor(10*np.random.random(3,4))
>>> a
array([[ 2., 8., 0., 6.],
[ 4., 5., 1., 1.],
[ 8., 9., 3., 6.]])
>>> a.shape
(3, 4)
>>> a.ravel() # returns the array, flattened
array([ 2., 8., 0., 6., 4., 5., 1., 1., 8., 9., 3., 6.])
>>> a.reshape(6,2) # returns the array with a modified shape
array([[ 2., 8.],
[ 0., 6.],
[ 4., 5.],
[ 1., 1.],
[ 8., 9.],
[ 3., 6.]])
>>> a.T # returns the array, transposed
array([[ 2., 4., 8.],
[ 8., 5., 9.],
[ 0., 1., 3.],
[ 6., 1., 6.]])
>>> a.T.shape
(4, 3)
>>> a.shape
(3, 4)
reshape返回重构后的数组，不改变引用的数组本身，而resize直接在引用的数组上重构。
>>> a
array([[ 2., 8., 0., 6.],
[ 4., 5., 1., 1.],
[ 8., 9., 3., 6.]])
>>> a.resize((2,6))
>>> a
array([[ 2., 8., 0., 6., 4., 5.],
[ 1., 1., 8., 9., 3., 6.]])
如果整形操作中指定尺寸为-1,则自动计算尺寸
>>> a.reshape(3,-1)
array([[ 2., 8., 0., 6.],
[ 4., 5., 1., 1.],
[ 8., 9., 3., 6.]])
堆叠
多个数组堆叠到一起
>>> a = np.floor(10*np.random.random(2,2))
>>> a
array([[ 8., 8.],
[ 0., 0.]])
>>> b = np.floor(10*np.random.random(2,2))
>>> b
array([[ 1., 8.],
[ 0., 4.]])
>>> np.vstack((a,b))
array([[ 8., 8.],
[ 0., 0.],
[ 1., 8.],
[ 0., 4.]])
>>> np.hstack((a,b))
array([[ 8., 8., 1., 8.],
[ 0., 0., 0., 4.]])
column_stack是将一维数组作为列形成二维数组，
>>> from numpy import newaxis
>>> np.column_stack((a,b)) # With 2D arrays
array([[ 8., 8., 1., 8.],
[ 0., 0., 0., 4.]])
>>> a = np.array([4.,2.])
>>> b = np.array([2.,8.])
>>> a[:,newaxis] # This allows to have a 2D columns vector
array([[ 4.],
[ 2.]])
>>> np.column_stack((a[:,newaxis],b[:,newaxis]))
array([[ 4., 2.],
[ 2., 8.]])
>>> np.vstack((a[:,newaxis],b[:,newaxis])) # The behavior of vstack is different
array([[ 4.],
[ 2.],
[ 2.],
[ 8.]])
如果想用更复杂的堆叠可以使用c_和r_,他们允许使用":"
>>> np.r_[1:4,0,4]
array([1, 2, 3, 0, 4])
一个数组切分成多个小片段数组
使用hsplit可以指定拆分的数组个数，或者在指定横轴位置后拆分
>> a = np.floor(10*np.random.random(2,12))
>>> a
array([[ 9., 5., 6., 3., 6., 8., 0., 7., 9., 7., 2., 7.],
[ 1., 4., 9., 2., 2., 1., 0., 6., 2., 2., 4., 0.]])
>>> np.hsplit(a,3) # 拆分成等份的三个数组
[array([[ 9., 5., 6., 3.],
[ 1., 4., 9., 2.]]), array([[ 6., 8., 0., 7.],
[ 2., 1., 0., 6.]]), array([[ 9., 7., 2., 7.],
[ 2., 2., 4., 0.]])]
>>> np.hsplit(a,(3,4)) # 在第三列后和第四列后拆分
[array([[ 9., 5., 6.],
[ 1., 4., 9.]]), array([[ 3.],
[ 2.]]), array([[ 6., 8., 0., 7., 9., 7., 2., 7.],
[ 2., 1., 0., 6., 2., 2., 4., 0.]])]
复制和视图
简单的赋值并不会产生复制对象
>>> a = np.arange(12)
>>> b = a # no new object is created
>>> b is a # a and b are two names for the same ndarray object
True
>>> b.shape = 3,4 # changes the shape of a
>>> a.shape
(3, 4)
>>> def f(x):
... print(id(x))
...
>>> id(a) # id is a unique identifier of an object
148293216
>>> f(a)
148293216
view or shadow copy
id不同但是改变元素值，二者都改变
>>> c = a.view()
>>> c is a
False
>>> c.base is a # c is a view of the data owned by a
True
>>> c.flags.owndata
False
>>>
>>> c.shape = 2,6 # a's shape doesn't change
>>> a.shape
(3, 4)
>>> c[0,4] = 1234 # a's data changes
>>> a
array([[ 0, 1, 2, 3],
[1234, 5, 6, 7],
[ 8, 9, 10, 11]])
分片产生的是视图（view）
>>> s = a[ : , 1:3] # spaces added for clarity; could also be written "s = a[:,1:3]"
>>> s[:] = 10 # s[:] is a view of s. Note the difference between s=10 and s[:]=10
>>> a
array([[ 0, 10, 10, 3],
[1234, 10, 10, 7],
[ 8, 10, 10, 11]])
深度复制
>>> d = a.copy() # a new array object with new data is created
>>> d is a
False
>>> d.base is a # d doesn't share anything with a
False
>>> d[0,0] = 9999
>>> a
array([[ 0, 10, 10, 3],
[1234, 10, 10, 7],
[ 8, 10, 10, 11]])
广播规则
广播规则允许universal function 在有意义的方法处理不同shape的输入数据。
当两个数组运算时，numpy逐个比较他们的shape，从最后一个维度比较，并且比较过程在方法工作的前面
可以广播的规则如下
- 他们是相等的
- 或者其中一个为1
否则返回ValueError: frames are not aligned
例子：
3和3相等所以可以广播，结果为3
Image (3d array): 256 x 256 x 3
Scale (1d array): 3
Result (3d array): 256 x 256 x 3
1和5中有个1,可以广播，结果返回5
A (4d array): 8 x 1 x 6 x 1
B (3d array): 7 x 1 x 5
Result (4d array): 8 x 7 x 6 x 5
更多的例子
A (2d array): 5 x 4
B (1d array): 1
Result (2d array): 5 x 4
A (2d array): 5 x 4
B (1d array): 4
Result (2d array): 5 x 4
A (3d array): 15 x 3 x 5
B (3d array): 15 x 1 x 5
Result (3d array): 15 x 3 x 5
A (3d array): 15 x 3 x 5
B (2d array): 3 x 5
Result (3d array): 15 x 3 x 5
A (3d array): 15 x 3 x 5
B (2d array): 3 x 1
Result (3d array): 15 x 3 x 5
>>> x = np.arange(4)
>>> xx = x.reshape(4,1)
>>> y = np.ones(5)
>>> z = np.ones((3,4))
>>> x.shape
(4,)
>>> y.shape
(5,)
>>> x + y
<type 'exceptions.ValueError'>: shape mismatch: objects cannot be broadcast to a single shape
>>> xx.shape
(4, 1)
>>> y.shape
(5,)
>>> (xx + y).shape
(4, 5)
>>> xx + y
array([[ 1., 1., 1., 1., 1.],
[ 2., 2., 2., 2., 2.],
[ 3., 3., 3., 3., 3.],
[ 4., 4., 4., 4., 4.]])
>>> x.shape
(4,)
>>> z.shape
(3, 4)
>>> (x + z).shape
(3, 4)
>>> x + z
array([[ 1., 2., 3., 4.],
[ 1., 2., 3., 4.],
[ 1., 2., 3., 4.]])
花式索引和索引技巧
NumPy提供比常规Python序列更多的索引功能。除了通过整数和切片进行索引之外，正如我们之前所看到的，数组可以通过整数数组和布尔数组索引。
索引数组索引
>>> a = np.arange(12)**2 # the first 12 square numbers
>>> i = np.array( [ 1,1,3,8,5 ] ) # an array of indices
>>> a[i] # the elements of a at the positions i
array([ 1, 1, 9, 64, 25])
>>>
>>> j = np.array( [ [ 3, 4], [ 9, 7 ] ] ) # a bidimensional array of indices
>>> a[j] # the same shape as j
array([[ 9, 16],
[81, 49]])
当数组a为多维数组时，索引指向a的第一维度，下面为一个多维数组的例子
>>> palette = np.array( [ [0,0,0], # black
... [255,0,0], # red
... [0,255,0], # green
... [0,0,255], # blue
... [255,255,255] ] ) # white
>>> image = np.array( [ [ 0, 1, 2, 0 ], # each value corresponds to a color in the palette
... [ 0, 3, 4, 0 ] ] )
>>> palette[image] # the (2,4,3) color image
array([[[ 0, 0, 0],
[255, 0, 0],
[ 0, 255, 0],
[ 0, 0, 0]],
[[ 0, 0, 0],
[ 0, 0, 255],
[255, 255, 255],
[ 0, 0, 0]]])
我们也可以指定超过一个索引，但是索引数组一定是相同的shape
>>> a = np.arange(12).reshape(3,4)
>>> a
array([[ 0, 1, 2, 3],
[ 4, 5, 6, 7],
[ 8, 9, 10, 11]])
>>> i = np.array( [ [0,1], # indices for the first dim of a
... [1,2] ] )
>>> j = np.array( [ [2,1], # indices for the second dim
... [3,3] ] )
>>>
>>> a[i,j] # i and j must have equal shape
array([[ 2, 5],
[ 7, 11]])
>>>
>>> a[i,2]
array([[ 2, 6],
[ 6, 10]])
>>>
>>> a[:,j] # i.e., a[ : , j]
array([[[ 2, 1],
[ 3, 3]],
[[ 6, 5],
[ 7, 7]],
[[10, 9],
[11, 11]]])
自然的，我们可以把i，j放到一个list中，然后使用这个list来索引
>>> l = [i,j]
>>> a[l] # equivalent to a[i,j]
array([[ 2, 5],
[ 7, 11]])
然而，我们不能把i，j放到一个数组中然后索引，会造成i，j都指向第一维度的索引，而出现超出索引范围的编译错误
>>> s = np.array( [i,j] )
>>> a[s] # not what we want
Traceback (most recent call last):
File "<stdin>", line 1, in ?
IndexError: index (3) out of range (0<=index<=2) in dimension 0
>>>
>>> a[tuple(s)] # same as a[i,j]
array([[ 2, 5],
[ 7, 11]])
另一个使用是查询时间系数序列的最大值
>>> time = np.linspace(20, 145, 5) # time scale
>>> data = np.sin(np.arange(20)).reshape(5,4) # 4 time-dependent series
>>> time
array([ 20. , 51.25, 82.5 , 113.75, 145. ])
>>> data
array([[ 0. , 0.84147098, 0.90929743, 0.14112001],
[-0.7568025 , -0.95892427, -0.2794155 , 0.6569866 ],
[ 0.98935825, 0.41211849, -0.54402111, -0.99999021],
[-0.53657292, 0.42016704, 0.99060736, 0.65028784],
[-0.28790332, -0.96139749, -0.75098725, 0.14987721]])
>>>
>>> ind = data.argmax(axis=0) # index of the maxima for each series
>>> ind
array([2, 0, 3, 1])
>>>
>>> time_max = time[ ind] # times corresponding to the maxima
>>>
>>> data_max = data[ind, xrange(data.shape[1])] # => data[ind[0],0], data[ind[1],1]...
>>>
>>> time_max
array([ 82.5 , 20. , 113.75, 51.25])
>>> data_max
array([ 0.98935825, 0.84147098, 0.99060736, 0.6569866 ])
>>>
>>> np.all(data_max == data.max(axis=0))
True
也可以使用它，对数组赋值
>>> a = np.arange(5)
>>> a
array([0, 1, 2, 3, 4])
>>> a[[1,3,4]] = 0
>>> a
array([0, 0, 2, 0, 0])
布尔数组索引
当我们用（整数）索引数组索引数组时，我们提供了要选择的索引列表。布尔指数的方法是不同的; 我们明确地选择我们想要的数组中的哪些项目，哪些不是我们想要的。
布尔索引最自然的方法就是使用与原始数组shape相同的布尔数组：
>>> a = np.arange(12).reshape(3,4)
>>> b = a > 4
>>> b # b is a boolean with a's shape
array([[False, False, False, False],
[False, True, True, True],
[ True, True, True, True]], dtype=bool)
>>> a[b] # 1d array with the selected elements
array([ 5, 6, 7, 8, 9, 10, 11])
这个方法在赋值中很有用，省略了循环结构
>>> a[b] = 0 # All elements of 'a' higher than 4 become 0
>>> a
array([[0, 1, 2, 3],
[4, 0, 0, 0],
[0, 0, 0, 0]])
另一种使用方法和数字索引很相似，对于每一个维度使用一个一维度bool数组来选择我们想要的分片
>>> a = np.arange(12).reshape(3,4)
>>> b1 = np.array([False,True,True]) # first dim selection
>>> b2 = np.array([True,False,True,False]) # second dim selection
>>>
>>> a[b1,:] # selecting rows
array([[ 4, 5, 6, 7],
[ 8, 9, 10, 11]])
>>>
>>> a[b1] # same thing
array([[ 4, 5, 6, 7],
[ 8, 9, 10, 11]])
>>>
>>> a[:,b2] # selecting columns
array([[ 0, 2],
[ 4, 6],
[ 8, 10]])
>>>
>>> a[b1,b2] # a weird thing to do
array([ 4, 10])
ix_()函数
网格化函数,和ogrid，mgrid，meshgrid一类，通过这个函数再结合广播规则。
ix_()函数可用于不同的载体结合，以便获得对于每一个n-uplet结果。
>>> a = np.arange(10).reshape(2, 5)
>>> a
array([[0, 1, 2, 3, 4],
[5, 6, 7, 8, 9]])
>>> ixgrid = np.ix_([0,1], [2,4])
>>> ixgrid
(array([[0],
[1]]), array([[2, 4]]))
>>> ixgrid[0].shape, ixgrid[1].shape
((2, 1), (1, 2))
>>> a[ixgrid]
array([[2, 4],
[7, 9]])
例如，如果要计算从每个矢量a，b和c取得的所有三元组的所有a + b * c：
>>> a = np.array([2,3,4,5])
>>> b = np.array([8,5,4])
>>> c = np.array([5,4,6,8,3])
>>> ax,bx,cx = np.ix_(a,b,c)
>>> ax
array([[[2]],
[[3]],
[[4]],
[[5]]])
>>> bx
array([[[8],
[5],
[4]]])
>>> cx
array([[[5, 4, 6, 8, 3]]])
>>> ax.shape, bx.shape, cx.shape
((4, 1, 1), (1, 3, 1), (1, 1, 5))
>>> result = ax+bx*cx
>>> result
array([[[42, 34, 50, 66, 26],
[27, 22, 32, 42, 17],
[22, 18, 26, 34, 14]],
[[43, 35, 51, 67, 27],
[28, 23, 33, 43, 18],
[23, 19, 27, 35, 15]],
[[44, 36, 52, 68, 28],
[29, 24, 34, 44, 19],
[24, 20, 28, 36, 16]],
[[45, 37, 53, 69, 29],
[30, 25, 35, 45, 20],
[25, 21, 29, 37, 17]]])
>>> result[3,2,4]
17
>>> a[3]+b[2]*c[4]
17
线性代数
linalg.py包
>>> import numpy as np
>>> a = np.array([[1.0, 2.0], [3.0, 4.0]])
>>> print(a)
[[ 1. 2.]
[ 3. 4.]]
>>> a.transpose()
array([[ 1., 3.],
[ 2., 4.]])
>>> np.linalg.inv(a)
array([[-2. , 1. ],
[ 1.5, -0.5]])
>>> u = np.eye(2) # unit 2x2 matrix; "eye" represents "I"
>>> u
array([[ 1., 0.],
[ 0., 1.]])
>>> j = np.array([[0.0, -1.0], [1.0, 0.0]])
>>> np.dot (j, j) # matrix product
array([[-1., 0.],
[ 0., -1.]])
>>> np.trace(u) # trace
2.0
>>> y = np.array([[5.], [7.]])
>>> np.linalg.solve(a, y)
array([[-3.],
[ 4.]])
>>> np.linalg.eig(j)
(array([ 0.+1.j, 0.-1.j]), array([[ 0.70710678+0.j , 0.70710678-0.j ],
[ 0.00000000-0.70710678j, 0.00000000+0.70710678j]]))
Parameters:
square matrix
Returns
The eigenvalues, each repeated according to its multiplicity.
The normalized (unit "length") eigenvectors, such that the
column ``v[:,i]`` is the eigenvector corresponding to the
eigenvalue ``w[i]`` .
技巧和提示
给出一个简短而有用的提示列表。
自动reshape
>>> a = np.arange(30)
>>> a.shape = 2,-1,3 # -1 means "whatever is needed"
>>> a.shape
(2, 5, 3)
>>> a
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]]])
数组堆叠
与matlab方式m[x;y](x,y都为相同长度矩阵)相同 ，
python可以使用column_stack, dstack, hstack and vstack
x = np.arange(0,10,2) # x=([0,2,4,6,8])
y = np.arange(5) # y=([0,1,2,3,4])
m = np.vstack([x,y]) # m=([[0,2,4,6,8],
# [0,1,2,3,4]])
xy = np.hstack([x,y]) # xy =([0,2,4,6,8,0,1,2,3,4])
histgram
应用于数组的NumPy 直方图函数返回一对向量：数组的直方图和分箱的向量。当心： matplotlib也有一个函数来建立直方图（称为hist，如在Matlab中），不同于NumPy中的。主要区别是pylab.hist自动绘制直方图，而 numpy.histogram只生成数据。
>>> import numpy as np
>>> import matplotlib.pyplot as plt
>>> # Build a vector of 10000 normal deviates with variance 0.5^2 and mean 2
>>> mu, sigma = 2, 0.5
>>> v = np.random.normal(mu,sigma,10000)
>>> # Plot a normalized histogram with 50 bins
>>> plt.hist(v, bins=50, normed=1) # matplotlib version (plot)
>>> plt.show()
>>> # Compute the histogram with numpy and then plot it
>>> (n, bins) = np.histogram(v, bins=50, normed=True) # NumPy version (no plot)
>>> plt.plot(.5*(bins[1:]+bins[:-1]), n)
>>> plt.show()
