Title: Indexing And Slicing Numpy Arrays
Slug: indexing_and_slicing_numpy_arrays
Summary: Indexing And Slicing Numpy Arrays
Date: 2016-05-01 12:00
Category: Python
Tags: Basics
Authors: Chris Albon

Slicing Arrays

Explanation Of Broadcasting

Unlike many other data types, slicing an array into a new variable means that any chances to that new variable are broadcasted to the original variable. Put other way, a slice is a hotlink to the original array variable, not a seperate and independent copy of it.



In [1]:

    
# Import Modules
import numpy as np



In [2]:

    
# Create an array of battle casualties from the first to the last battle
battleDeaths = np.array([1245, 2732, 3853, 4824, 5292, 6184, 7282, 81393, 932, 10834])



In [3]:

    
# Divide the array of battle deaths into start, middle, and end of the war
warStart = battleDeaths[0:3]; print('Death from battles at the start of war:', warStart)
warMiddle = battleDeaths[3:7]; print('Death from battles at the middle of war:', warMiddle)
warEnd = battleDeaths[7:10]; print('Death from battles at the end of war:', warEnd)









    



Death from battles at the start of war: [1245 2732 3853]
Death from battles at the middle of war: [4824 5292 6184 7282]
Death from battles at the end of war: [81393   932 10834]



In [4]:

    
# Change the battle death numbers from the first battle
warStart[0] = 11101



In [5]:

    
# View that change reflected in the warStart slice of the battleDeaths array
warStart









    Out[5]:





array([11101,  2732,  3853])



In [6]:

    
# View that change reflected in (i.e. "broadcasted to) the original battleDeaths array
battleDeaths









    Out[6]:





array([11101,  2732,  3853,  4824,  5292,  6184,  7282, 81393,   932, 10834])

Indexing Arrays

Note: This multidimensional array behaves like a dataframe or matrix (i.e. columns and rows)



In [7]:

    
# Create an array of regiment information
regimentNames = ['Nighthawks', 'Sky Warriors', 'Rough Riders', 'New Birds']
regimentNumber = [1, 2, 3, 4]
regimentSize = [1092, 2039, 3011, 4099]
regimentCommander = ['Mitchell', 'Blackthorn', 'Baker', 'Miller']

regiments = np.array([regimentNames, regimentNumber, regimentSize, regimentCommander])
regiments









    Out[7]:





array([['Nighthawks', 'Sky Warriors', 'Rough Riders', 'New Birds'],
       ['1', '2', '3', '4'],
       ['1092', '2039', '3011', '4099'],
       ['Mitchell', 'Blackthorn', 'Baker', 'Miller']], 
      dtype='<U12')



In [8]:

    
# View the first column of the matrix
regiments[:,0]









    Out[8]:





array(['Nighthawks', '1', '1092', 'Mitchell'], 
      dtype='<U12')



In [9]:

    
# View the second row of the matrix
regiments[1,]









    Out[9]:





array(['1', '2', '3', '4'], 
      dtype='<U12')



In [10]:

    
# View the top-right quarter of the matrix
regiments[:2,2:]









    Out[10]:





array([['Rough Riders', 'New Birds'],
       ['3', '4']], 
      dtype='<U12')