How can I preserve line endings when reading and writing a text file?

Use newline='' when opening the copied sample for both reading and writing. In text mode, Python’s default newline=None recognizes \n, \r, and \r\n on input but returns each as \n. That makes different original endings indistinguishable in the text you read. With newline='', Python still recognizes those line boundaries but returns their ending characters untranslated. On output, newline=None translates written \n characters to the system line separator; newline='' disables that translation. [S1]

The setting does not stop your own code from changing the text. If you call a method that normalizes line breaks, or rebuild the file using a different separator, the original pattern can still be lost. Work on a copy, retain the ending characters, and check the result.

Here is a minimal copy-based example. Replace the bracketed filenames and choose an encoding appropriate for your sample. Encoding and newline handling are separate settings; this example does not determine which encoding your file uses. [S1]

from pathlib import Path

source = Path("[copied sample path]")
destination = Path("[output path]")

with source.open("r", encoding="[appropriate encoding]", newline="") as f:
    text = f.read()

# Make any intended edits without normalizing line endings.

with destination.open("w", encoding="[appropriate encoding]", newline="") as f:
    f.write(text)

For a preservation test with no intended edits, leave text untouched. The output is a separate file, so you can compare it with the copied input without overwriting the original.

Inspect the text using repr() so line-ending characters are visible rather than appearing as ordinary layout. For example, the illustrative string "first\r\nsecond\nthird\r" contains three different endings. You can count the characters in the read text like this:

print(repr(text))
print("CRLF:", text.count("\r\n"))
print("LF:", text.count("\n") - text.count("\r\n"))
print("CR:", text.count("\r") - text.count("\r\n"))

Run the same check on the copied input and the output, reading each with newline=''. Matching visible text and counts is a useful check for this example, but it is not a byte-for-byte comparison: text decoding and re-encoding can affect bytes, and this method does not establish that every other file property stayed unchanged. For byte-level control, Python’s documentation points to binary I/O. [S1]

Make the narrow edit on the text you read with newline='', then write that text using newline='' as well. A simple string replacement can illustrate the approach, but first ensure the target is the intended text and that your change does not include or replace line-ending characters:

text = text.replace("[old word]", "[new word]")

Avoid reading with the default newline handling when you need to distinguish the original endings. Avoid writing with newline=None when the text contains line breaks you intend to leave untranslated. And do not assume the sample’s encoding: use a suitable, known encoding for both operations. The documentation explains newline translation and notes that encoding must also be selected appropriately; it does not identify the encoding of your particular file or guarantee preservation of all file bytes through text I/O. [S1]