数据类型（dtype）¶

文档时效性说明

本文为早期笔记，可能存在版本过时、命令失效、链接失效、最佳实践变化等问题。请以官方最新文档为准。

原英文标题：Data Types (dtpye) #

bool
int
float
complex

1. Arrays¶

数组可以通过列表、元组或函数生成。

1.1 使用列表生成数组¶

numpy.array(object, dtype=None, copy=True, order=None, subok=False, ndmin=0)

我们只需要关注 dtype 参数。它可以是 int、float、complex、bool...

注意：如果不指定 dtype，系统会自动选择合适的类型

示例：

np.array([(1, 2), (3, 4), (5, 6)])

1.2 使用函数生成数组¶

1.2.1 np.arange()¶

1
2
3

*示例*
``` python
np.arange(3, 7, 0.5, dtype='float32')

输出

array([3. , 3.5, 4. , 4.5, 5. , 5.5, 6. , 6.5])

np.arange(number)

示例

np.arrange(10)

输出

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

np.linspace()¶

numpy.linspace(start, stop, num=50, endpoint=True, retstep=False, dtype=None)

示例（endpoint=True）

np.linspace(0, 10, 10, endpoint=True)

输出

array([ 0.        ,  1.11111111,  2.22222222,  3.33333333,  4.44444444,
        5.55555556,  6.66666667,  7.77777778,  8.88888889, 10.        ])

示例（endpoint=False）

np.linspace(0, 10, 10, endpoint=False)

输出

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

np.ones()¶

np.ones 用于创建所有元素均为 1 的数组。

numpy.ones(shape, dtype=None, order='C')

示例

np.ones((2, 3))

输出

array([[1., 1., 1.],
       [1., 1., 1.]])

注意其中的 "2" 称为 "axis 0"，而 "3" 称为 "axis 1"。

在二维数组中，"axis 0" 是列，"axis 1" 是行。

np.zeros()¶

numpy.zeros(shape, dtype=None, order='C')

示例

np.zeros((3, 2))

输出

array([[0., 0.],
       [0., 0.],
       [0., 0.]])

np.eye()¶

numpy.eye() 创建一个对角线上为 1、其余位置为 0 的数组。

numpy.eye(N, M=None, k=0, dtype=<type 'float'>)

其中 k 表示对角线偏移量。N 定义列上的元素数量，M 定义行上的元素数量。M 的默认值等于 N。

见以下 3 个示例。

示例

np.eye(5)

输出

array([[1., 0., 0., 0., 0.],
       [0., 1., 0., 0., 0.],
       [0., 0., 1., 0., 0.],
       [0., 0., 0., 1., 0.],
       [0., 0., 0., 0., 1.]])

示例

np.eye(5,3)

输出

array([[1., 0., 0.],
       [0., 1., 0.],
       [0., 0., 1.],
       [0., 0., 0.],
       [0., 0., 0.]])

示例

np.eye(5, 3, -2)

输出

array([[0., 0., 0.],
       [0., 0., 0.],
       [1., 0., 0.],
       [0., 1., 0.],
       [0., 0., 1.]])

np.fromfunction()¶

示例

np.fromfunction(lambda a, b: a + b, (5, 4))

输出

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

注意行列的索引从 0 开始，而不是 1。

数组操作¶

设置数组的数据类型¶

a.astype(int)

获取数组的类型¶

a.dtype

转置数组¶

a.T

或者使用 transpose 函数

a = np.arange(4).reshape(2, 2)
np.transpose(a)

array([[0, 2],
       [1, 3]])

获取实部和虚部¶

a.real
a.imag

获取大小、形状和维度¶

a.size
a.ndim
a.shape

重塑和 Resize¶

重塑 (Reshape)¶

np.reshape(newshape, order='C')

示例

a=np.arange(10)
a.reshape((5, 2))

输出

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

示例

np.arange(10).reshape((5, 2), order='F')

输出

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

调整大小 (Resize)¶

np.resize(a，new_shape)

示例

a = np.arange(10)
a.resize(2, 5)
a

输出

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

展平 (Ravel)¶

np.ravel(array, order='C')

示例

np.ravel(a)

输出

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

示例

np.ravel(a, order='F')

输出

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

改变轴¶

np.moveaxis(a, source, destination)

示例

a = np.ones((1, 2, 3))
print(a)
np.moveaxis(a, 0, -1)

输出

[[[1. 1. 1.]
  [1. 1. 1.]]]

array([[[1.],
        [1.],
        [1.]],

       [[1.],
        [1.],
        [1.]]])

注意在三维数组中，"axis 0" 代表高度，"axis 1" 和 "axis 2" 分别代表 "列" 和 "行"

np.swapaxis(a, axis1, axis2)

示例

a = np.ones((1, 4, 3))
print(a)
np.swapaxes(a, 0, 2)

输出

[[[1. 1. 1.]
  [1. 1. 1.]
  [1. 1. 1.]
  [1. 1. 1.]]]

array([[[1.],
        [1.],
        [1.],
        [1.]],

       [[1.],
        [1.],
        [1.],
        [1.]],

       [[1.],
        [1.],
        [1.],
        [1.]]])

改变维度¶

np.atleast_1d()
np.atleast_2d()
np.atleast_3d()

示例

print(np.atleast_1d([1, 2, 3]))
print(np.atleast_2d([4, 5, 6]))
print(np.atleast_3d([7, 8, 9]))

输出

[1 2 3]
[[4 5 6]]
[[[7]
   [8]
   [9]]]

拼接 (Concatenate)¶

np.concatenate((a1, a2, ...), axis=0)

示例

a = np.array([[1, 2], [3, 4], [5, 6]])
b = np.array([[7, 8], [9, 10]])
c = np.array([[11, 12]])

np.concatenate((a, b, c), axis=0)

输出

array([[ 1,  2],
       [ 3,  4],
       [ 5,  6],
       [ 7,  8],
       [ 9, 10],
       [11, 12]])

示例

a = np.array([[1, 2], [3, 4], [5, 6]])
b = np.array([[7, 8, 9]])

np.concatenate((a, b.T), axis=1)

输出

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

拆分 (Split)¶

示例

a = np.arange(10)
np.split(a, 5)

输出

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

示例

a = np.arange(10).reshape(2, 5)
np.split(a, 2)

输出

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

删除 (Delete)¶

np.delete(arr，obj，axis)

示例

a = np.arange(12).reshape(3, 4)
np.delete(a, 2, 1)

输出

array([[ 0,  1,  3],
       [ 4,  5,  7],
       [ 8,  9, 11]])

插入 (Insert)¶

np.insert(arr，obj，values，axis)

示例

a = np.arange(12).reshape(3, 4)
b = np.arange(4)

np.insert(a, 2, b, 0)

输出

array([[ 0,  1,  2,  3],
       [ 4,  5,  6,  7],
       [ 0,  1,  2,  3],
       [ 8,  9, 10, 11]])

追加 (Append)¶

np.append(arr，values，axis)

示例

a = np.arange(6).reshape(2, 3)
b = np.arange(3)

np.append(a, b)

输出

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

翻转 (Flipping)¶

a = np.arange(16).reshape(4, 4)
print(np.fliplr(a))
print(np.flipud(a))

[[ 3  2  1  0]
 [ 7  6  5  4]
 [11 10  9  8]
 [15 14 13 12]]
[[12 13 14 15]
 [ 8  9 10 11]
 [ 4  5  6  7]
 [ 0  1  2  3]]

Markdown Code

随机数组¶

np.random.rand(2, 5)

np.random.rand(2, 5)

array([[0.09433914, 0.08680661, 0.23040579, 0.71954424, 0.54292341],
       [0.22890897, 0.49553437, 0.01181691, 0.10668025, 0.71153526]])

np.random.randint(2, 5, 10)

array([3, 3, 4, 4, 2, 4, 4, 2, 4, 2])

np.random.random_sample([10])

array([0.80117316, 0.48038627, 0.40861977, 0.22925529, 0.91899056,
       0.70100459, 0.21080387, 0.94939295, 0.374128  , 0.28534828])

均匀分布 (Uniform Distribution)¶

np.random.rand(shape)

正态分布 (Normal Distribution)¶

np.random.randn(shape)

学生分布 (Student Distribution)¶

numpy.random.standard_t(df，size)

其他分布¶

    numpy.random.beta(a，b，size)
    numpy.random.binomial(n, p, size)
    numpy.random.chisquare(df，size)
    numpy.random.dirichlet(alpha，size)
    numpy.random.exponential(scale，size)
    numpy.random.f(dfnum，dfden，size)
    numpy.random.gamma(shape，scale，size)
    numpy.random.geometric(p，size)
    numpy.random.gumbel(loc，scale，size)
    numpy.random.hypergeometric(ngood, nbad, nsample, size)
    numpy.random.laplace(loc，scale，size)
    numpy.random.logistic(loc，scale，size)
    numpy.random.lognormal(mean，sigma，size)
    numpy.random.logseries(p，size)
    numpy.random.multinomial(n，pvals，size)
    numpy.random.multivariate_normal(mean, cov, size)
    numpy.random.negative_binomial(n, p, size)
    numpy.random.noncentral_chisquare(df，nonc，size)
    numpy.random.noncentral_f(dfnum, dfden, nonc, size)
    numpy.random.normal(loc，scale，size)
    numpy.random.pareto(a，size)
    numpy.random.poisson(lam，size)    numpy.random.standard_exponential(size)
    numpy.random.standard_gamma(shape，size)
    numpy.random.standard_normal(size)
    numpy.random.standard_t(df，size)
    numpy.random.triangular(left，mode，right，size)
    numpy.random.uniform(low，high，size)
    numpy.random.vonmises(mu，kappa，size)
    numpy.random.wald(mean，scale，size)
    numpy.random.weibull(a，size)
    numpy.random.zipf(a，size)

函数¶

原文（English）¶

---
tags:
  - Python
---

# 2. 数据类型（dtype）


!!! warning "文档时效性说明"
    本文为早期笔记，可能存在版本过时、命令失效、链接失效、最佳实践变化等问题。请以官方最新文档为准。

* bool
* int
* float
* complex

# 3. Arrays #

Arrays may be generated by lists or tuples, or functions.

## 3.1 Using lists to generate arrays ##

```python
numpy.array(object, dtype=None, copy=True, order=None, subok=False, ndmin=0)

All we need to consider is the dtype parameter. It may be int, float, complex, bool...

Note that dtype can be selected automatically if you do not specify one

Example:

np.array([(1, 2), (3, 4), (5, 6)])

3.2 Use Functions to generate arrays¶

3.2.1 np.arange()¶

1
2
3

*Example*
``` python
np.arange(3, 7, 0.5, dtype='float32')

Output

array([3. , 3.5, 4. , 4.5, 5. , 5.5, 6. , 6.5])

np.arange(number)

Example

np.arrange(10)

Output

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

np.linspace()¶

numpy.linspace(start, stop, num=50, endpoint=True, retstep=False, dtype=None)

Example (endpoint=True)

np.linspace(0, 10, 10, endpoint=True)

Output

array([ 0.        ,  1.11111111,  2.22222222,  3.33333333,  4.44444444,
        5.55555556,  6.66666667,  7.77777778,  8.88888889, 10.        ])

Example (endpoint=False)

np.linspace(0, 10, 10, endpoint=False)

Output

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

np.ones()¶

np.ones is used for creating arrays whose elements are all 1.

numpy.ones(shape, dtype=None, order='C')

Example

np.ones((2, 3))

Output

array([[1., 1., 1.],
       [1., 1., 1.]])

Note that the "2" is called "axis 0", while the "3" is called "axis 1".

In 2 dimension arrays, "axis 0" is the column, "axis 1" is the line.

np.zeros()¶

numpy.zeros(shape, dtype=None, order='C')

Example

np.zeros((3, 2))

Output

array([[0., 0.],
       [0., 0.],
       [0., 0.]])

np.eye()¶

numpy.eye() creates an array which has value 1 on its diagonal and 0 on other positions.

numpy.eye(N, M=None, k=0, dtype=<type 'float'>)

Whereas k means the offset of diagonal. N defines the amount of elements on the column, M defines the amount of elements on the row. The default value of M is equal to N.

See the 3 examples below.

Example

np.eye(5)

Output

array([[1., 0., 0., 0., 0.],
       [0., 1., 0., 0., 0.],
       [0., 0., 1., 0., 0.],
       [0., 0., 0., 1., 0.],
       [0., 0., 0., 0., 1.]])

Example

np.eye(5,3)

Output

array([[1., 0., 0.],
       [0., 1., 0.],
       [0., 0., 1.],
       [0., 0., 0.],
       [0., 0., 0.]])

Example

np.eye(5, 3, -2)

Output

array([[0., 0., 0.],
       [0., 0., 0.],
       [1., 0., 0.],
       [0., 1., 0.],
       [0., 0., 1.]])

np.fromfunction()¶

Example

np.fromfunction(lambda a, b: a + b, (5, 4))

Output

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

Notes that the index of column and row counts from 0, not 1.

Operating with arrays¶

Set the data type of an array¶

a.astype(int)

Get the type of an array¶

a.dtype

Transpose an array¶

a.T

or you may use the transpose function

a = np.arange(4).reshape(2, 2)
np.transpose(a)

array([[0, 2],
       [1, 3]])

Get the real and imaginary part¶

a.real
a.imag

Get size, shape and dimension¶

a.size
a.ndim
a.shape

Reshape and Resize¶

Reshape¶

np.reshape(newshape, order='C')

Example

a=np.arange(10)
a.reshape((5, 2))

Output

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

Example

np.arange(10).reshape((5, 2), order='F')

Output

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

Resize¶

np.resize(a，new_shape)

Example

a = np.arange(10)
a.resize(2, 5)
a

Output

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

Ravel¶

np.ravel(array, order='C')

Example

np.ravel(a)

Output

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

Example

np.ravel(a, order='F')

Output

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

Change axis¶

np.moveaxis(a, source, destination)

Example

a = np.ones((1, 2, 3))
print(a)
np.moveaxis(a, 0, -1)

Output

[[[1. 1. 1.]
  [1. 1. 1.]]]

array([[[1.],
        [1.],
        [1.]],

       [[1.],
        [1.],
        [1.]]])

Note that in a three dimension array, "axis 0" represents the height, "axis 1" and "axis 2" represents of "column" and "row"

np.swapaxis(a, axis1, axis2)

Example

a = np.ones((1, 4, 3))
print(a)
np.swapaxes(a, 0, 2)

Output

[[[1. 1. 1.]
  [1. 1. 1.]
  [1. 1. 1.]
  [1. 1. 1.]]]

array([[[1.],
        [1.],
        [1.],
        [1.]],

       [[1.],
        [1.],
        [1.],
        [1.]],

       [[1.],
        [1.],
        [1.],
        [1.]]])

Change the dimension¶

np.atleast_1d()
np.atleast_2d()
np.atleast_3d()

Example

print(np.atleast_1d([1, 2, 3]))
print(np.atleast_2d([4, 5, 6]))
print(np.atleast_3d([7, 8, 9]))

Output

[1 2 3]
[[4 5 6]]
[[[7]
   [8]
   [9]]]

Concatenate¶

np.concatenate((a1, a2, ...), axis=0)

Example

a = np.array([[1, 2], [3, 4], [5, 6]])
b = np.array([[7, 8], [9, 10]])
c = np.array([[11, 12]])

np.concatenate((a, b, c), axis=0)

Output

array([[ 1,  2],
       [ 3,  4],
       [ 5,  6],
       [ 7,  8],
       [ 9, 10],
       [11, 12]])

Example

a = np.array([[1, 2], [3, 4], [5, 6]])
b = np.array([[7, 8, 9]])

np.concatenate((a, b.T), axis=1)

Output

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

Split¶

Example

a = np.arange(10)
np.split(a, 5)

Output

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

Example

a = np.arange(10).reshape(2, 5)
np.split(a, 2)

Output

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

Delete¶

np.delete(arr，obj，axis)

Example

a = np.arange(12).reshape(3, 4)
np.delete(a, 2, 1)

Output

array([[ 0,  1,  3],
       [ 4,  5,  7],
       [ 8,  9, 11]])

Insert¶

np.insert(arr，obj，values，axis)

Example

a = np.arange(12).reshape(3, 4)
b = np.arange(4)

np.insert(a, 2, b, 0)

Output

array([[ 0,  1,  2,  3],
       [ 4,  5,  6,  7],
       [ 0,  1,  2,  3],
       [ 8,  9, 10, 11]])

Append¶

np.append(arr，values，axis)

Example

a = np.arange(6).reshape(2, 3)
b = np.arange(3)

np.append(a, b)

Output

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

Flipping¶

a = np.arange(16).reshape(4, 4)
print(np.fliplr(a))
print(np.flipud(a))

[[ 3  2  1  0]
 [ 7  6  5  4]
 [11 10  9  8]
 [15 14 13 12]]
[[12 13 14 15]
 [ 8  9 10 11]
 [ 4  5  6  7]
 [ 0  1  2  3]]

Markdown Code

Random arrays¶

np.random.rand(2, 5)

np.random.rand(2, 5)

array([[0.09433914, 0.08680661, 0.23040579, 0.71954424, 0.54292341],
       [0.22890897, 0.49553437, 0.01181691, 0.10668025, 0.71153526]])

np.random.randint(2, 5, 10)

array([3, 3, 4, 4, 2, 4, 4, 2, 4, 2])

np.random.random_sample([10])

array([0.80117316, 0.48038627, 0.40861977, 0.22925529, 0.91899056,
       0.70100459, 0.21080387, 0.94939295, 0.374128  , 0.28534828])

Uniform Distribution¶

np.random.rand(shape)

Normal Distribution¶

np.random.randn(shape)

Student Distribution¶

numpy.random.standard_t(df，size)

Other Distributions¶

    numpy.random.beta(a，b，size)
    numpy.random.binomial(n, p, size)
    numpy.random.chisquare(df，size)
    numpy.random.dirichlet(alpha，size)
    numpy.random.exponential(scale，size)
    numpy.random.f(dfnum，dfden，size)
    numpy.random.gamma(shape，scale，size)
    numpy.random.geometric(p，size)
    numpy.random.gumbel(loc，scale，size)
    numpy.random.hypergeometric(ngood, nbad, nsample, size)
    numpy.random.laplace(loc，scale，size)
    numpy.random.logistic(loc，scale，size)
    numpy.random.lognormal(mean，sigma，size)
    numpy.random.logseries(p，size)
    numpy.random.multinomial(n，pvals，size)
    numpy.random.multivariate_normal(mean, cov, size)
    numpy.random.negative_binomial(n, p, size)
    numpy.random.noncentral_chisquare(df，nonc，size)
    numpy.random.noncentral_f(dfnum, dfden, nonc, size)
    numpy.random.normal(loc，scale，size)
    numpy.random.pareto(a，size)
    numpy.random.poisson(lam，size)    numpy.random.standard_exponential(size)
    numpy.random.standard_gamma(shape，size)
    numpy.random.standard_normal(size)
    numpy.random.standard_t(df，size)
    numpy.random.triangular(left，mode，right，size)
    numpy.random.uniform(low，high，size)
    numpy.random.vonmises(mu，kappa，size)
    numpy.random.wald(mean，scale，size)
    numpy.random.weibull(a，size)
    numpy.random.zipf(a，size)

Functions¶

```