文本文件无法在AWS Elastic Beanstalk（Flask Web App）上正确显示

Question

I have created a machine learning web app and I converted the dataframe predictions into a .txt file that users can download

df.to_csv('pred.txt', sep=',', index=False)

and the output format should be like (on localhost Windows 10):

x,y,z
1,2,3
4,5,6
7,8,9

However, when I deployed my web app on AWS Elastic Beanstalk (linux) and download the pred.txt, the format became like:

x,y,z,1,2,3,4,5,6,7,8,9

Why is the format like that and how should I change it so that the format is like the the prior format?

Answer 1

Windows and Linux have different newline characters:

https://en.wikipedia.org/wiki/Newline#Representation

You can force to_csv() to use the Windows newlines by providing it as line_terminator , like so:

df.to_csv('pred.txt', sep=',', index=False, line_terminator='\\r\\n')