This section covers how to read from files in Python and how to write to them.

You need three steps to read text from a file:

  • open the file,
  • read something from it,
  • and finally close it.

Create a poem.txt file in the editor and write any poem into it. Save the file.

In the following example, we are using encoding utf-8, which covers majority of european characters.

There are other encodings covering different character sets. If you want to learn more about encodings, you can read this nicely readable Encoding article

Let's create a following code:

poem_file = open('poem.txt', encoding='utf-8')
content =

Run it in the directory with poem.txt. In other words, the current working directory must contain the file with the poem.

The program prints the poem!

What's going on here? The open() function returns the value that represents the open file. This value has its own methods. We are using the read() method that reads the entire contents of the file at once and returns it as a string. We will cover close(), the function that closes the open file, later.

Iteration over contents of files

You can use open files with the for statement. It's similar as with strings or ranges. for i in range provides consecutive numbers. for c in 'abcd' provides single string characters. for line in poem_file: provides individual lines read from the file into the line variable.

For example, we can indent the poem to make it stand out of the text.

print('I heard this poem:')
poem_file = open('poem.txt', encoding='utf-8')
for line in poem_file:
    print('    ' + line)
print('How do you like it?')

When you try it, you will find that the spacing is not how it should be. Would you like to try to explain why this is so?


Closing files

It is quite important to close the file after the program stops using it. The close() method does this for us. Operating systems have limits on open files. If you do not close them you can exceed this limit. Besides, on Windows, you cannot re-open a file that is already open.

You can compare files to a fridge: If you want to put something into the fridge, you need to open it and then close it. The fridge works without closing, too, but then something goes rotten.

It is easy to forget to close a file. For example, an exception or return statement inside the file processing may skip the close(). Then the file remains open.

We can use the try/finally statement to make sure that the file is closed.

The finally block (the statements(s) after finally) is always executed. It executes no matter if the try blocks ends with success, or with an exception, or if you jump out of it using return or break.

def initial_character():
    """Return the first character in the poem."""

    poem_file = open('poem.txt', encoding='utf-8')
        content =
        return content[0]


You can use the finally block every time you need to close or terminate something -- not just a file, it can also be a database connection.

The with statement

Because the try/finally block is quite verbose, there is a better way in Python. It's the with statement, that we have seen during the negative testing using pytest (with pytest.raises()):

def initial_character():
    """Return the first character in the poem."""

    with open('poem.txt', encoding='utf-8') as poem_file:
        content =
        return content[0]


We used this statement for testing before. It wraps a block with an expected exception. It checks if the correct exception has occurred after the block ends. In our case, the file is closed when the block ends no matter what has happened. The file is closed in all cases if the with block ends with success, or with an exception, or if we're jumping out of it.

The with statement is the best option for working with files in the majority of cases.

Writing to files


It is easy to delete or overwrite any file in Python. Try the following examples in a directory where you have nothing important!

You can write to a file in Python. You need to open the file for writing using a named argument mode='w' (w stands for write). You can write individual strings using the write() method.

If the file already exists, opening it with mode='w' overwrites its original content. There will be only the text that your program writes into it.

with open('second-poem.txt', mode='w', encoding='utf-8') as poem_file:
    poem_file.write('Our old chiming clock\n')
    poem_file.write("Is beating four o'clock\n")

Why is there a `\n`?

The write() method does not put a line ending after the string. If you need to write multiple lines to files, you need to
end each of them by a newline character '\n'. We have described it in the Strings section.

Or, you can use the print() function. By default, it writes to the terminal. It can also write into an open file if you use the named argument file.

Other print() options remain unchanged. These options include line ending, conversion to strings, and printing multiple arguments at a time.

with open('second-poem.txt', mode='w', encoding='utf-8') as poem_file:
    print('Our old chiming clock', file=poem_file)
    print('Is beating', 2+2, "o'clock", file=poem_file)

Writing and reading JSON data

JSON can be stored into .txt files as well. It it done in a similar way, we just need to import the json library.

If we want to store our data from a dictionary into the file, we can use following code:

data = {
    "name": "Anna",
    "city": "Brno",
    "languages": ["Czech", "English", "Python"],
    "age": 26

import json

with open('json_data.txt', mode='w', encoding='utf-8') as json_file:
    json.dump(data, json_file)
import json

with open('json_data.txt') as json_file:
    data = json.load(json_file)


In previous lesson json we have shown also loading JSON from a string and writing to a string, which used methods loads and dumps. The ending s represents string. You can remember the difference between load and loads like that.

Other common data formats

In order to read or write csv you can use the built-in csv library.

For other data formats like Excel files, you need to install external libraries.

For working with tabular data it is the easiest to use pandas library - which also needs to be installed extra.