5 Best Ways to Convert a Python NumPy Array to a Table

by

💡 Problem Formulation: Data scientists and developers often need to present numerical data stored in a NumPy array in a tabular format for analysis, reporting, or debugging purposes. The challenge is converting a 2D NumPy array, which might look something like numpy.array([[1, 2], [3, 4]]), into a readable table representation that resembles: | 1 | 2 | | 3 | 4 | This article demonstrates five methods for making this conversion smoothly.

Method 1: Using Pandas DataFrame

Pandas is a powerful data manipulation library. Converting a NumPy array to a table can be done seamlessly with Pandas by creating a DataFrame. The DataFrame constructor accepts a NumPy array as an input and automatically formats it into a table-like object.

Here’s an example:

import numpy as np
import pandas as pd

# Creating NumPy array
array = np.array([[1, 2], [3, 4]])

# Converting to Pandas DataFrame
df_table = pd.DataFrame(array)

# Displaying the table
print(df_table)

Output of this code snippet:

   0  1
0  1  2
1  3  4

This code snippet creates a NumPy array and then uses Pandas to cast it into a DataFrame, which inherently displays data in a tabular form. Note that each column and row gets automatically labeled with an index.

Method 2: Using PrettyTable Library

The PrettyTable library offers a way to generate ASCII tables from NumPy arrays. It’s ideal for quick visualization in terminal applications and supports many formatting options to customize the table’s appearance.

Here’s an example:

from prettytable import PrettyTable
import numpy as np

# Creating NumPy array
array = np.array([[1, 2], [3, 4]])

# Creating a PrettyTable
table = PrettyTable()

# Specifying column names if desired
table.field_names = ["Column 1", "Column 2"]

# Adding rows to the PrettyTable
for row in array:
    table.add_row(row)

print(table)

Output of this code snippet:

+----------+----------+
| Column 1 | Column 2 |
+----------+----------+
|    1     |    2     |
|    3     |    4     |
+----------+----------+

This snippet creates an ASCII table from a NumPy array using PrettyTable by iterating over the array rows and adding them as rows to the table.

Method 3: Using tabulate Library

The tabulate library provides an easy way to render a NumPy array as a table in various output formats such as plain-text, HTML, LaTeX, and more. It is simple, intuitive, and requires only a couple of lines to achieve the desired table.

Here’s an example:

from tabulate import tabulate
import numpy as np

# Creating NumPy array
array = np.array([[1, 2], [3, 4]])

# Using tabulate to display the table
print(tabulate(array, headers="firstrow", tablefmt="grid"))

Output of this code snippet:

+---+---+
| 1 | 2 |
+---+---+
| 3 | 4 |
+---+---+

The code uses the tabulate function to convert an array into a grid-formatted table. The headers and tablefmt parameters define how the table headers and overall format should appear.

Method 4: Using CSV Writer

If the table needs to be persisted to a file format, using Python’s csv module can be a direct approach. It allows for writing NumPy arrays to a CSV file, which is a standard format for storing tabular data.

Here’s an example:

import numpy as np
import csv

# Creating NumPy array
array = np.array([[1, 2], [3, 4]])

# Writing array to a CSV file
with open("data.csv", "w", newline="") as csvfile:
    writer = csv.writer(csvfile)
    writer.writerows(array)

In this example, the array data is written to a CSV file named ‘data.csv’. This file can then be opened by any spreadsheet software and will be displayed as a table.

Bonus One-Liner Method 5: Using NumPy’s savetxt

NumPy has built-in functions for file I/O. The savetxt function can write an array to a file in a table-like format, including custom delimiters and headers if needed.

Here’s an example:

import numpy as np

# Creating NumPy array
array = np.array([[1, 2], [3, 4]])

# Writing array to a text file
np.savetxt("data.txt", array, fmt="%d", delimiter="\t", header="Column1\tColumn2")

The savetxt function writes the array to ‘data.txt’, using a tab as a delimiter between columns, and adds a header row at the top.

Summary/Discussion

  • Method 1: Pandas DataFrame. Strengths: Provides a full-fledged DataFrame object with extensive manipulation capabilities. Weaknesses: Requires Pandas library which is large and not necessary if just a basic table representation is required.
  • Method 2: PrettyTable Library. Strengths: Aesthetically pleasing ASCII tables with high customizability. Weaknesses: Externally dependent library and unsuitable for very large data sets.
  • Method 3: Tabulate Library. Strengths: Highly versatile in terms of the output formats and yet simple to use. Weaknesses: Requires installation of an external package.
  • Method 4: CSV Writer. Strengths: Native to Python; outputs in a universal format (CSV) which can be easily shared and viewed. Weaknesses: Not immediately viewable in script output; must be opened with another application.
  • Method 5: NumPy’s savetxt. Strengths: Part of NumPy, avoiding external dependencies; one-liner solution. Weaknesses: Limited to simple text files and lacks the robust features of other CSV-specific methods.