NumPy Foundation (i)

Install self-addressed

# #为什么使用NumPy

File vectorsumcompare.py

#!/usr/bin/env python# -*- coding:utf-8 -*-__author__ =  ' Teng ' import  Sysfrom datetime import datetimeimport numpy as npdef numpysum (n):     a = np.arange (n) **2    b = np.arange (n) **3     c = a+b    return cdef pythonsum (n):     a = range (n)     b = range (n)     c  = []    for i in range (Len (a)):         a[i] = i**2        b[i] = i**3         c.append (A[i]+ b[i])     return  Csize = int (Sys.argv[1]) Start = datetime.now () c = pythonsum (size) print  " Pythonsum: ",  cdelta = dateTime.now ()  - startprint  "The last 2 elements of the sum",  c[ -2:]print  "Pythonsum elapsed time in microseconds",  delta.microsecondsstart  = datetime.now () c = numpysum (size) print  "Numpysum:", cdelta =  DateTime.Now ()  - startprint  "The last 2 elements of the sum",  c[-2:]print  "Numpysum elapsed time in microseconds",  delta.microseconds

Run the above script like Python vectorsumcompare.py 10000

Advantages of NumPy

Simple

Fast speed when the data volume is large

# #NumPy数组对象

Debug method shape Returns the size of an element in a tuple tuple to each dimension of the NumPy array

Arange one-dimensional arrays

in [+]: M = Np.array ([Np.arange (2), Np.arange (2)])

in [+]: M

OUT[16]: Array ([[[0, 1],[0, 1]])

in [+]: M.shape

OUT[17]: (2, 2)

Ndarray is a multidimensional array object:

is divided into two parts of the actual data and the metadata that describes the data

NumPy Foundation (i)