How do I choose and specify a text encoding when opening a file in Python?

Choose the encoding based on information about how the file was produced—not merely on what you hope it uses. If the source documentation or export settings say the file is UTF-8, for example, pass that value explicitly:

with open("sample.txt", "r", encoding="utf-8") as file:
    text = file.read()

The encoding argument tells open() which encoding to use for text. If you leave it out, Python uses a platform-dependent encoding, so the same program can behave differently across environments. [S1]

Treat utf-8 here as an example, not a guarantee about your file. Replace it with the encoding supported by evidence about that file. Keeping the choice explicit also makes the assumption visible to whoever reads or runs the code.

Do not count on automatic identification being reliable. The Python documentation supports supplying an encoding, but does not promise that an unknown file’s encoding can always be inferred. [S1]

Work from the file’s history instead: ask which application created it, check any export options or accompanying documentation, and find out whether the sender can confirm the setting. If you have multiple files from the same source, verify whether they were produced using the same settings rather than assuming they were.

If you still have to investigate, try a plausible candidate on a copy of the file and examine the decoded text, especially places where you expect non-English characters or symbols. A result that looks readable is a clue, not proof that you identified the original encoding. Record the candidate and how you selected it so the assumption is easy to revisit.

Open the original for reading and inspect a small portion of the resulting text; this example does not write to or modify the file:

with open("sample.txt", "r", encoding="utf-8") as file:
    preview = file.read(500)

print(preview)

The number 500 is just an example for limiting this inspection; it is not a rule about encodings. Try a different candidate only when you have a reason to consider it, and compare the parts of the text that matter. For instance, if you expect names or accented letters, check whether those characters appear as expected.

For a more reusable test, keep the candidate in a clearly named variable and include it in your notes:

candidate_encoding = "utf-8"

with open("sample.txt", "r", encoding=candidate_encoding) as file:
    preview = file.read(500)

This makes the choice easy to see and change. Looking at a preview can help you assess a candidate, but it does not independently establish the file’s original encoding.

Keep a short note alongside the script or file-handling instructions. For example:

File: [filename or file pattern]
Encoding used: [encoding]
Reason: [documentation, export setting, or sender confirmation]
Checked: [what text or characters you inspected]
Remaining uncertainty: [anything not confirmed]

If the encoding is confirmed by the file’s producer or documentation, say that. If you chose it by inspecting a preview, describe it as a tested candidate instead. The distinction matters: Python lets you specify an encoding, but the fact that a program opens a file does not, by itself, verify that your choice matches how the file was created. [S1]