Recently, I was working on a data analysis project where I needed to export NumPy arrays to text files that could be shared with team members who didn’t use Python. The solution? NumPy’s efficient np.savetxt() function.
In this article, I’ll show you everything you need to know about using np.savetxt() in Python to save your NumPy arrays to text files. I’ll cover the basic syntax, formatting options, and share some practical examples that you can use in your projects.
So let’s get in!
np.savetxt() Function in Python
The np.savetxt() function is a NumPy utility that allows you to save array data to a text file. It’s particularly useful when you need to:
- Export data for use in other applications like Excel or MATLAB
- Share results with colleagues who don’t use Python
- Create CSV files from your NumPy arrays
- Archive numerical data in a human-readable format
Read Python NumPy Matrix Operations
Basic Syntax and Parameters
Here’s the basic syntax of the np.savetxt() function:
np.savetxt(fname, X, fmt='%.18e', delimiter=' ', newline='\n', header='', footer='', comments='# ', encoding=None)Let’s explore the key parameters:
fname: The filename or file handle where data will be savedX: The array data to be savedfmt: Format specifier for the datadelimiter: String that separates values in the output fileheader: String to be written at the beginning of the filefooter: String to be written at the end of the filecomments: String that will be prepended to the header and footer
Check out 0-Dimensional Array NumPy in Python
Save NumPy Arrays to Text Files in Python
Now, I will explain how to save NumPy arrays to text files in Python.
Method 1 – Simple Array Export with Default Settings
The simplest way to use np.savetxt() is to provide just the filename and the Python array:
import numpy as np
# Create structured array
population_data = np.array([
[39.5, "California"],
[30.0, "Texas"],
[21.8, "Florida"],
[19.8, "New York"],
[13.0, "Pennsylvania"]
], dtype=object)
# Print to console
print("Population (millions)\tState")
for row in population_data:
print(f"{row[0]}\t{row[1]}")
# Save to text file
np.savetxt(
'us_population.txt',
population_data,
fmt='%s\t%s',
header='Population (millions)\tState',
comments=''
)Output:
Population (millions) State
39.5 California
30.0 Texas
21.8 Florida
19.8 New York
13.0 PennsylvaniaYou can refer to the screenshot below to see the output.
However, this will attempt to convert all elements to float, which can cause issues with mixed data types. Let’s look at a better way to handle this.
Method 2 – Format Output with the fmt Parameter
When working with specific data types or when you want to control how your data appears in the output file, the fmt parameter is your friend:
import numpy as np
import os
# Create numeric data - US GDP by quarter for 2022 (in trillion $)
gdp_data = np.array([
[2022, 1, 24.386],
[2022, 2, 24.978],
[2022, 3, 25.723],
[2022, 4, 26.138]
])
# Ensure the directory exists
directory = r'C:\Users\Public\code'
if not os.path.exists(directory):
os.makedirs(directory)
# Define the file path
file_path = os.path.join(directory, 'us_gdp_2022.csv')
# Save with formatted output
np.savetxt(file_path,
gdp_data,
fmt='%d,%d,%.3f', # Integer year, integer quarter, 3 decimal places for GDP
header='Year,Quarter,GDP_Trillion',
comments='') # Empty comments to avoid # before header
print(f"File saved to {file_path}")Output:
File saved to C:\Users\Public\code\us_gdp_2022.csvYou can refer to the screenshot below to see the output.
This creates a clean CSV file with properly formatted numbers and a header line.
Check out Create an Empty Array using NumPy in Python
Method 3 – Work with CSV Files using Delimiters
CSV files are one of the most common formats for data exchange. Here’s how to create them with np.savetxt():
import numpy as np
# US stock market data (date, open, high, low, close)
stock_data = np.array([
['2023-01-05', 3808.10, 3822.93, 3802.43, 3808.10],
['2023-01-06', 3807.40, 3906.19, 3807.40, 3895.08],
['2023-01-09', 3892.38, 3950.57, 3891.68, 3892.09],
['2023-01-10', 3888.57, 3919.83, 3877.29, 3919.25]
])
# Save as CSV with appropriate formatting
np.savetxt('sp500_data.csv',
stock_data,
fmt='%s,%.2f,%.2f,%.2f,%.2f',
delimiter=',', # Using comma as delimiter
header='Date,Open,High,Low,Close',
comments='')The delimiter parameter ensures values are separated correctly, while the fmt parameter handles the mixed data types.
Read NumPy: Create a NaN Array in Python
Method 4 – Add Headers and Footers
For better documentation of your data, you can add headers and footers:
import numpy as np
# US temperature data for major cities (city, temp in F)
temp_data = np.array([
[98.6, 102.3, 97.1, 99.5],
[97.8, 101.2, 98.6, 100.1],
[98.1, 103.4, 96.8, 98.7]
])
# Save with headers and footers
np.savetxt('patient_temperatures.csv',
temp_data,
fmt='%.1f',
delimiter=',',
header='New York,Los Angeles,Chicago,Houston',
footer='Data collected on 2023-01-15',
comments='# ')The comments parameter (#) adds a hash before the header and footer lines, marking them as comments in many data processing tools.
Check out NumPy Zeros in Python
Method 5 – Handle Complex Data Types
When working with complex data types, you might need to convert your data:
import numpy as np
# Complex data with mixed types
complex_data = np.array([
(1, 'John', 'CEO', 250000),
(2, 'Lisa', 'CTO', 225000),
(3, 'Mike', 'CFO', 215000)
], dtype=[('id', 'i4'), ('name', 'U10'), ('position', 'U10'), ('salary', 'f8')])
# Convert structured array to a simpler format for savetxt
formatted_data = np.array([
[row['id'], row['name'], row['position'], row['salary']]
for row in complex_data
])
# Save to file
np.savetxt('employee_data.csv',
formatted_data,
fmt='%d,%s,%s,%.2f',
delimiter=',',
header='ID,Name,Position,Salary',
comments='')By converting the structured array to a simpler format first, we make it easier to save with np.savetxt().
Check out NumPy Read CSV with Header in Python
Method 6 – Load the Saved Data Back into Python
To complete the cycle, you can load the data back using np.loadtxt():
import numpy as np
# First, save some US sales data
sales_data = np.array([
[1, 45000, 52000, 47000],
[2, 42000, 49000, 44000],
[3, 50000, 55000, 52000],
[4, 47000, 51000, 49000]
])
np.savetxt('quarterly_sales.csv',
sales_data,
fmt='%d,%.2f,%.2f,%.2f',
delimiter=',',
header='Quarter,East,West,Central',
comments='')
# Now load it back
loaded_data = np.loadtxt('quarterly_sales.csv',
delimiter=',',
skiprows=1) # Skip the header row
print("Loaded data:", loaded_data)The skiprows parameter in np.loadtxt() allows us to skip the header when loading the data back.
I explained six methods in this tutorial to save NumPy arrays to text files that are array exporting with default settings, formatting output with the fmt parameter, working with CSV files using a delimiter, handling complex data types, and loading the saved data back into Python.
Other Python articles you may also like:
I am Bijay Kumar, a Microsoft MVP in SharePoint. Apart from SharePoint, I started working on Python, Machine learning, and artificial intelligence for the last 5 years. During this time I got expertise in various Python libraries also like Tkinter, Pandas, NumPy, Turtle, Django, Matplotlib, Tensorflow, Scipy, Scikit-Learn, etc… for various clients in the United States, Canada, the United Kingdom, Australia, New Zealand, etc. Check out my profile.
