Numpy quick start (I) -- pre knowledge (create array + constant + data type)

One 、 What is? Numpy？

NumPy yes Python The basic package of scientific computing in . It's a Python library , Provide multidimensional array objects , Derived objects and various... For quick array operations API, Including mathematics 、 Logic 、 Shape operation 、 Sort 、 choice 、 Input and output 、 Basic linear algebra , Basic statistical operation and random simulation, etc .

NumPy At the heart of yes ndarray（n-dimensional array） object , namely n Dimension group , It is associated with Python The differences between lists are as follows ：

• NumPy Arrays are created with a fixed size , and Python The list of can dynamically grow or shorten . change ndarray The size of will create a new array and delete the original array .
• NumPy Elements in an array need to have the same data type , and Python The list of can have elements of different data types .
• NumPy Arrays help with advanced mathematics and other types of operations on large amounts of data , Usually these operations are performed more efficiently , Than using Python The list has less code .
• More and more based on Python Science of / Math package use NumPy Array , Although these tools usually support Python As a parameter , But they will still convert the input list into NumPy Array of , And it's usually output as NumPy Array .

Two 、 install Numpy

stay Command line Use the following commands to install ：

pip install numpy

about conda user , Available

conda install numpy

After the installation , We can import Numpy 了 , We usually use the following code to import ：

import numpy as np

np yes numpy Abbreviation , The purpose is to make the program more concise and convenient , Use at the same time np Will not lose readability , And this abbreviation is also the choice of the vast majority of people .

3、 ... and 、ndarray and matrix Which one to use ？

We already know ,ndarray yes Numpy Medium n Dimensional array object , But beyond that ,Numpy A special matrix type is also provided ：matrix, It is ndarray Of Subclass .

ndarray and matrix Can complete the operation of matrix , So which one is better ？

Let's start with the answer ： Use ndarray.

The reasons are as follows ：

• ndarray yes Numpy The standard vector in / matrix / Tensor type , many Numpy function It's an array , Not the matrix .
• Use ndarray Than using matrix It is more convenient to build a matrix .
• ndarray Can handle high-dimensional arrays , and matrix Only two-dimensional arrays can be processed .
• stay ndarray in , Operator * Represents multiplying by elements （ It can also be used. multiply()）, Operator @ Representation matrix multiplication （ It can also be used. dot()）; And in the matrix in ,* It represents matrix multiplication , Multiplying by elements requires multiply().

Four 、 establish n Dimension group

You must have been right about Numpy Medium n Dimensional array object ndarray Have a general understanding , So how do we create n And the dimension group ？

There are two common methods ：

• Put the existing Python list 、 Tuple to ndarray
• Use Numpy The built-in array creation function （ Such as arange、linspace etc. ）

4.1 Put the existing list 、 Tuple to n Dimension group

It's easy , Just put one on the outside np.array() that will do .

For a list of ：

A = np.array([1, 2, 3])
print(A)
print(type(A))
# [1 2 3]
# <class 'numpy.ndarray'>

For tuples :

A = np.array((1, 2, 3))
print(A)
print(type(A))
# [1 2 3]
# <class 'numpy.ndarray'>

4.2 Use Numpy The built-in array creation function

We will only make a brief introduction here , About the creation of arrays 、 See the next article in this series for details of modifications and so on .

function	describe
np.arange([start], stop, [step])	Create one in [start, stop) In the interval , In steps of step Array of ;start When omitted, the default is 0,step When omitted, the default is 1
np.linspace(start, stop, num)	Create one in [start, stop] In the interval , The size is num Array of

for example , To create $[0,10]$ All even arrays in , We can use arange function , Just set the step size to 2：

A = np.arange(0, 11, 2)
print(A)
# [ 0 2 4 6 8 10]

You can also use linspace function , But we need to know in advance $[0,10]$ How many even numbers in ：

A = np.linspace(0, 10, 6)
print(A)
# [ 0. 2. 4. 6. 8. 10.]

look arange and linspace The result is the same , But the reality is different , I won't go into details here , See the next article for details .

4.3 Yes ndarray Index and slice

Of course , We can treat as Python Like the list , Yes Numpy Medium n Dimension groups are indexed and sliced ：

A = np.arange(10)
print(A[3])
print(A[1:])
print(A[:-2])
print(A[::-1])
print(A[::2])
# 3
# [1 2 3 4 5 6 7 8 9]
# [0 1 2 3 4 5 6 7]
# [9 8 7 6 5 4 3 2 1 0]
# [0 2 4 6 8]

5、 ... and 、Numpy Constant in

Constant	describe
np.inf	representative $+\infty$
np.NINF	representative $-\infty$
np.nan	For non numeric
np.e	Represents the natural base
np.pi	representative $\pi$
np.euler_gamma	Represents Euler constant
np.newaxis	representative None

Be careful ： As mentioned above 7 Of the constants , The first six are all floating-point , namely <class 'float'>, Only the last one is Nonetype type ：<class 'NoneType'>.

The first six constants must be familiar to everyone , And it is often used in practical programming tasks . For the seventh constant ：np.newaxis, We will explain it in detail in the following article .

6、 ... and 、Numpy Data types in

Numpy Support than Python More types of data , This section will cover that common Data type of , And how to modify n Data type of dimension group .

6.1 Boolean type

name	describe
np.bool8（np.bool_）	Boolean data type （True perhaps False）

6.2 integer

name	describe
np.int8	byte （-128 to 127）
np.int16	Integers （-32768 to 32767）
np.int32	Integers （-2147483648 to 2147483647）
np.int64（np.int_）	Integers （-9223372036854775808 to 9223372036854775807）
np.uint8	Unsigned integer （0 to 255）
np.uint16	Unsigned integer （0 to 65535）
np.uint32	Unsigned integer （0 to 4294967295）
np.uint64	Unsigned integer （0 to 18446744073709551615）

6.3 floating-point

name	describe
np.float16	Semi precision floating point number , Include ：1 Sign bits ,5 One digit ,10 One last digit
np.float32	Single-precision floating-point , Include ：1 Sign bits ,8 One digit ,23 One last digit
np.float64（np.float_）	Double precision floating point , Include ：1 Sign bits ,11 One digit ,52 One last digit

6.4 Plural

name	describe
np.complex64	The plural , By two 32 Bit floating point （ Real and imaginary parts ） form
np.complex128（np.complex_）	The plural , By two 64 Bit floating point （ Real and imaginary parts ） form

6.5 dtype

6.5.1 View data type

Be careful ：Numpy The default data type in is float_, Double precision floating-point number

dtype Can be used to view ndarray The data type of the element in . Let's do some experiments first ：

Experiment 1

A = np.array([1, 1])
print(A.dtype)
# int32

Visible for n Dimensional integer array ,Numpy The default data type is int32.

in fact , If the value of an element in the array is greater than 2147483647 when , The data type will be converted to int64.

Experiment two

A = np.array([1., 1.])
print(A.dtype)
# float64

Visible for n Dimensional floating-point array ,Numpy The default data type is float_.

Experiment three

A = np.array([1j, 1j])
print(A.dtype)
# complex128

Visible for n Dimensional complex array ,Numpy The default data type is complex_.

6.5.2 Set the data type of the array

To set the data type of the array , We can do that ：

A = np.array([1, 2, 3], dtype=np.float64)
print(A)
# [1. 2. 3.]

You can find that integer arrays have been converted to floating-point arrays .

Of course, we have simpler methods ：

A = np.float64([1, 2, 3])
print(A)
# [1. 2. 3.]

The two methods mentioned above are both in Set the data type while creating the array , We can also use astype() Method Change the data type of the existing array ：

A = np.array([1, 2, 3])
print(A.dtype)
A = A.astype(np.float64)
print(A.dtype)
# int32
# float64

Most of the time , We all declare data types while creating arrays , In fact, we can use it more succinctly dtype：

A more concise form	Original form
dtype=int	Equivalent to dtype=np.int32
dtype=float	Equivalent to dtype=np.float64
dtype=complex	Equivalent to dtype=np.complex128

for example ：

A = np.array([1, 1], dtype=int)
print(A.dtype)
# int32

6.5.3 Spillover problem

If the data type is set improperly , May cause data overflow . for example , When creating an array like this ：

A = np.array([3333443333, 4444334444], dtype=np.int32)

Will report a mistake ： OverflowError, This is because the value of an element in the array is greater than int32 The scope of , We need to use int64 To continue creating .

A = np.array([3333443333, 4444334444], dtype=np.int64)
print(A)
# [3333443333 4444334444]