This is the first day of my participation in Gwen Challenge

The current plan for this series is to update pandas, and some specific cases will be explained together. Pandas will then compare the two technologies: SQL and PANDAS.

In this section, however, let’s review the common numpy. We’ll use it later in the code.

1. Numpy generation

  • The simplest way is to specify the value directly
import numpy as np
np.array([1.2.3])
Copy the code

Output: array([1, 2, 3])

  • Generate an arithmetic sequence

    • Contain termination

np.linspace(1.5.3) Start, end, array length
Copy the code

Array ([1., 3., 5.])

Note: these happen to be integers, but will be converted to floating point numbers.

    • Does not contain termination items
np.arange(1.5.2) Start, end, array length
Copy the code

Output: array([1, 3])

2. Special matrix

  • Np. zeros generate all zero matrices (probably not used much)

  • Np.eye (3) generates the identity matrix with 3 rows and 3 columns

  • Fill the matrix

np.full((6.5), 10)
Copy the code

np.full((6.5),1.2.3])

Copy the code

Np. Full ((6,5), [1, 2, 3, 4, 5])Copy the code

To sum up: The np.full function populates the columns based on their values. If the two parameter columns are not equal, they will not be filled unless the second parameter column is 1

3. Array deformation and merge

  • Array transpose
import numpy as np
arr = np.full((6, 5), [1, 2, 3, 4, 5])
print(arr)
arr.T
Copy the code
// The output is [[1 2 3 4 5] [1 2 3 4 5] [1 2 3 4 5] [1 2 3 4 5] [1 2 3 4 5] [1 2 3 4 5] [1 2 3 4 5] [1 2 3 4 5] [1 2 3 4 5] [1 2 3 4 5] [1 2 3 4 5]] array([[1, 1, 1, 1, 1], [2, 2, 2, 2, 1, 2, 1, 1, 1]) 2, 2, 2], [3, 3, 3, 3, 3, 3], [4, 4, 4, 4, 4, 4], [5, 5, 5, 5, 5, 5]])Copy the code
  • Array merge np.r_, np.c_
import numpy as np

print(np.r_[np.array([0.0]),np.zeros(2)])
print(np.c_[np.array([1.2]),np.zeros(2)])

print(np.r_[np.array([[0.0], [1.1]]),np.array([[2.2], [3.3]]])Copy the code

Look at the code above and the output is:

[0. 0 0. 0.] [[1. 0.] [2. 0.]] [[0, 0] [1, 1] [2] 2 3] [3]Copy the code

For one-dimensional arrays, the np.r_ and Np. c_ merge rules seem to be reversed. After repeated verification and analysis, it is found that numpy stores one-dimensional arrays as column vectors.