[Tut] How to Iterate over a NumPy Array

[xSicKxBot]

How to Iterate over a NumPy Array

Rate this post

Problem Formulation and Solution Overview

In this article, you’ll learn how to iterate over a 1D, 2D and 3D NumPy array using Python.

When working with the NumPy library, you will encounter situations where you will need to iterate through a 1D, 2D and even a 3D array. This article will show you how to accomplish this task.

Question: How would we write code to iterate through a [i]1D, 2D or 3D[/i] NumPy array?

We can accomplish this task by one of the following options:

• Method 1: Use a
  Code:

  For
  loop and
  Code:

  np.array()
  
  
  
• Method 2: Use a https://blog.finxter.com/for-else-and-wh...ful-guide/
  Code:

  For
  loop and
  Code:

  np.nditer()
  
  
  
• Method 3: Use a https://blog.finxter.com/for-else-and-wh...ful-guide/
  Code:

  For
  loop and
  Code:

  itertools
  
  
  
• Method 4: Use a
  Code:

  While
  loop and
  Code:

  np.size
  
  
  
• Method 5: Use a
  Code:

  For
  loop and
  Code:

  np.ndenumerate()
  
  
  
• Method 6: Use a
  Code:

  For
  Loop and
  Code:

  range()
  
  
  
• Bonus: CSV to
  Code:

  np.array()
  
  

Preparation

Before moving forward, please ensure the NumPy library is installed. Click here if you require instructions.

Then, add the following code to the top of each script. This snippet will allow the code in this article to run error-free.

Code:

import numpy as np

After importing the NumPy library, this library is referenced by calling the shortcode (

Code:

np

).

Method 1: Use a For Loop and np.array()

This method uses a

Code:

For

loop combined with

Code:

np.array()

to iterate through a 1D NumPy array. The first five (5) Atomic Numbers from the Periodic Table are generated and displayed for this example.

Code:

atomic_els = np.array(np.arange(1,6)) for el in atomic_els: print(el, end=' ')

Above, calls the

Code:

np.array()

function and passes it

Code:

np.arange(1,6)

with two (2) arguments: a start position of one (1) and stop position of five (5) or (stop-1). The result is an arrayhttps://blog.finxter.com/python-list-methods/of integers and saves to

Code:

atomic_els

.

Code:

[1 2 3 4 5]

Next, a

Code:

For

loop is instantiated and iterates through each element of the 1D NumPy array

Code:

atomic_els

.

The output is sent to the terminal on one (1) line as an additional argument was passed to the print statement (

Code:

end=' '

). This argument replaces the default newline character with a blank space.

Code:

1 2 3 4 5

Method 2: Uses a For Loop and nditer()

This method uses a https://blog.finxter.com/python-loops/

Code:

for

loop combined with https://blog.finxter.com/how-to-initiali...easy-ways/

Code:

np.nditer()

to iterate through a NumPy array. The first five (5) Atomic Numbers and Number of Neutrons data from the Periodic Table display for this example.

Code:

atomic_data = np.array([np.arange(1,6), [0, 2, 4, 5, 6]]) for dim in atomic_data: for d in dim: print(d, end=' ')

Above, calls the

Code:

np.array()

function and passes https://blog.finxter.com/python-range-fu...h%20Gauss.

Code:

np.arange(1,6)

as the first argument and an array of associated Number of [b]Neutrons[/b] values as the second.

Next, an outer

Code:

for

loop is instantiated. This iterates through each array dimension.

Inside this loop, another

Code:

for

loop is instantiated. The inside loop iterates through each element of the above dimensions.

The output is sent to the terminal on one (1) line as an additional argument is passed to the print statement (

Code:

end=' '

). This argument replaces the default newline character with a blank space.

Code:

1 2 3 4 5 0 2 4 5 6

Note: The first dimension is highlighted to easily differentiate the dimensions.



Method 3: Use a For Loop and itertools

This method uses a

Code:

for

loop and Python’s built-in

Code:

itertools

library to iterate through a NumPy array. The first three (3) Atomic Numbers, Phase, and Group data from the Periodic Table display for this example.

Code:

import itertools atomic_num = np.arange(1,4) atomic_phase = ['gas', 'gas', 'solid'] atomic_group = [1, 18, 1] for (a, b, c) in itertools.zip_longest(atomic_num, atomic_phase, atomic_group, fillvalue=0): print (a, b, c)

Above, imports the

Code:

itertools

library to use the

Code:

zip_longest()

function.

Then, three (3) arrays are declared containing the relevant data for the first three (3) elements from the Periodic Table. They save respectively to

Code:

atomic_num

,

Code:

atomic_phase

, and

Code:

atomic_group

.

Next, a

Code:

for

loop is instantiated and passed three (3) arguments inside the brackets

Code:

(a, b, c)

which reference the arguments inside the

Code:

zip_longest(atomic_num, atomic_phase, atomic_group)

function.

An additional argument (

Code:

fillvalue=0

) is also passed. This fills in any missing values for lists of unequal lengths with the stated value.

The loop iterates until the end of the arrays are reached, outputting to the terminal for each iteration.

Code:

1 gas 12 gas 183 solid 1

Method 4: Use a While Loop and Size

This method uses a

Code:

while

loop combined with https://blog.finxter.com/how-to-initiali...easy-ways/

Code:

np.size

to iterate through a NumPy array. The first five (5) Atomic Numbers, Number of Neutrons, and Phase data from the Periodic Table display for this example.

Code:

atomic_names = np.array(['Hydrogen', 'Helium', 'Lithium', 'Beryllium', 'Boron']) el_count = 0 while el_count < atomic_names.size: print(atomic_names[el_count], end=' ') el_count += 1

Above, calls the

Code:

np.array()

function and passes a list containing the first five (5) element names of the Periodic Table. These results save to

Code:

atomic_names

.

Then, a variable,

Code:

el_count

, is declared with a start value of zero (). This will be a counter variable and lets the

Code:

While

loop know when to stop iterating.

Next, a

Code:

while

loop is instantiated and iterates through the elements of

Code:

atomic_names

. This will continue while the value of

Code:

el_count

remains less than the value of

Code:

atomic_names.size

.

The output is sent to the terminal on one (1) line as an additional argument is passed to the print statement (

Code:

end=' '

). This argument replaces the default newline character with a blank space.

Code:

Hydrogen Helium Lithium Beryllium Boron

Method 5: Use a For Loop and np.ndenumerate()

This method uses a

Code:

For

loop and

Code:

np.ndenumerate()

to iterate through a NumPy array. The first five (5) Element Names from the Periodic Table display for this example.

Code:

atomic_names = np.array(['Hydrogen', 'Helium', 'Lithium', 'Beryllium', 'Boron']) for idx, x in np.ndenumerate(atomic_names): print(idx, x)

Above, calls the

Code:

np.array()

function and passes it an array containing the first five (5) element names from the Periodic Table. The results save to

Code:

atomic_names

.

Next, a

Code:

for

loop is instantiated referencing

Code:

idx

which is the index of the array, and

Code:

x

, which is the array element’s value.

The output is sent to the terminal. Then

Code:

idx

displays a

Code:

Tuple

containing the index value of the array and then the value of

Code:

x

for each iteration.

Code:

(0,) Hydrogen(1,) Helium(2,) Lithium(3,) Beryllium(4,) Boron

Method 6: Use a For Loop and range()

This method uses a

Code:

for

loop and

Code:

range()

) to iterate through a 3D NumPy array.

Code:

nums = np.array([[[1, 2], [3, 4], [5, 6]], [[7, 8], [9, 10], [11, 12]], [[13, 14], [15, 16], [17, 18]], [[19, 20], [21, 22], [23, 23]], [[24, 25], [26, 27], [28, 29]]]) for x in range(0, 5): for y in range(0, 3): for z in range(0, 2): print(nums[x][y][z])

Above declares a 3D NumPy array containing consecutive numbers from 1-29 inclusive. These save to

Code:

nums

.

Next, three (3)

Code:

for

loops are instantiated to loop through and output the contents of

Code:

nums

to the terminal one (1) number per line.

The range() function for each loop is based on the dimensions of the 3D Numpy array, for example:

• The first loop (
  Code:

  range(0,5
  ) identifies the total number of tows in the array.
  

• The next loop (
  Code:

  range(0,3
  ) identifies the number of columns in the array.
  

• The final loop (
  Code:

  range(0,2
  ) identifies the elements in each column in the array.
  

Finally, the output is sent to the terminal (snippet only).

Code:

12345678910...

Bonus: Convert CSV to np.array()

This example reads in a snippet of the Periodic Table as a CSV file. This data is converted to a NumPy array and is output to the terminal.

The CSV file below contains the values of the Atomic Number, Atomic Mass, Number of Neutrons, and the Number of Electrons for the first seven (7) elements from the Periodic Table.

Contents of

Code:

pt_sample.csv

Code:

1,1.007,02,4.002,23,6.941,44,9.012,55,10.811,66,12.011,67,14.007,7

To follow along, save this file as

Code:

pt_sample.csv

and move it into the current working directory.

Before moving forward, please ensure the Pandas library is installed. Click here if you require instructions.

Code:

import pandas as pd import csv file_name = 'pt_sample.csv' df_data = np.array(list(csv.reader(open(file_name, 'r'), delimiter=','))).astype("float") print(df_data)

Above, imports the Pandas library and the CSV library. This is needed to work with DataFrames and read in the CSV file.

Then, a filename is declared and saves to

Code:

file_name

.

On the highlighted line, the

Code:

np.array()

function is called and passed the following arguments.

• The
  Code:

  csv.reader()
  function is passed as an argument and this argument passes the
  Code:

  open()
  method to open the specified CSV file in read (
  Code:

  r
  ) mode, including the
  Code:

  field
  delimiter (
  Code:

  csv.reader(open(file_name, 'r'), delimiter=',')
  ).
  
  
• The contents of the CSV file converts to a list.
  
  
• This list converts to an
  Code:

  np.array()
  .
  

The results save to

Code:

df_data

and output to the terminal.

Code:

[[ 1. 1.007 0. ][ 2. 4.002 2. ][ 3. 6.941 4. ][ 4. 9.012 5. ][ 5. 10.811 6. ][ 6. 12.011 6. ][ 7. 14.007 7. ]]

Summary

These seven (7) methods of iterating a NumPy array should give you enough information to select the best one for your coding requirements.

Good Luck & Happy Coding!

Programmer Humor – Blockchain

[i]“Blockchains are like grappling hooks, in that it’s extremely cool when you encounter a problem for which they’re the right solution, but it happens way too rarely in real life.” source – xkcd[/i]