Read and write calls will fail at runtime if open is called with incorrect modes.

The second argument of open specifies the "mode" in which the file is opened. For example opening in write mode:

with open('some/path.txt', 'w') as f:
    f.write('foo')

If unspecified, mode defaults to 'rt' which allows calling read, readline and readlines. 't' means the contents will be of type str.

Python will throw an exception at runtime if a file read/write method is used without the file being opened in a mode that allows it. File modes are a source of bugs that will happen at runtime because there is no in-language feature preventing a developer from specifying modes that are invalid with the way the file will later be interacted with. If the invalid modes are not hit during testing then they will ultimately fail in prod.

The full options are:

In Python 2.7 'U' simplified the handling of newlines across platforms, so a file's new lines work on Unix even if created Windows. This is the default behaviour in Python 3, so 'U' is not needed in Python 3.

Reading

Python supports various methods of reading files: read, readline and readlines. To use the read methods 'r' mode must be in use (or must be in read/write mode):

with open('some/path.txt', 'r') as f:
    content = f.read()

with open('some/path.txt') as f:  # 'r' is the default mode
    content = f.read()

with open('some/path.txt') as f:
    content = f.readlines() # returns a list of lines

with open('some/path.txt') as f:
    line_one = f.readline() # returns one line at a time
    line_two = f.readline()
    line_three = f.readline()

Python will throw an exception if read, readline and readlines are used without the file being opened with 'r' mode:

with open('some/path.txt', 'w') as f:
    content = f.read()

UnsupportedOperation: not readable

Writing

Python supports write for writing to files. To use them 'w', or 'a', or 'x' mode must be in use (or read/write mode):

with open('some/path.txt', 'w') as f:  # open and truncate
    f.write('hello')

with open('some/path.txt', 'a') as f:  # open and not truncate
    f.write('hello')  # added to end of the existing contents

with open('some/path.txt', 'x') as f:  # open in exclusive mode
    f.write('hello')

Python will throw an exception if write is used without the file being opened with 'w', or 'a', or 'x' mode (or read/write mode):

with open('some/path.txt') as f:
    f.write('hello')

UnsupportedOperation: not writable

Python will also throw an exception if exclusive mode is used but the file already exists

with open('some/path.txt', 'x') as f:
    f.write('hello')

FileExistsError: [Errno 17] File exists

Exclusive mode is useful for creating thread-safe pid files:

try:
    with open("some/pid/file", "x") as f:
        f.write(...)
except FileExistsError:
    pass

Updating

Python allows the file to be opened in both read and write mode by using '+' alongside 'r', 'w', or 'a', so there is no need to open the same file multiple times in different modes.

Note though that 'r+', 'a+', and 'w+' do very different things:

• 'r+' opens an existing file for reading and writing without truncating it. The file pointer is at the start of the file, so read will return the current contents of the file, but write will override the current contents unless the file pointer is be moved to the end of the fie via f.seek(...)
• 'a+' opens a file for reading and writing without truncating it. The file pointer is at the end of the file, so read will return an '' (even if there is contents in the file), and write will append the end of the existing file contents. The file pointer can be moved to the start of the fie via f.seek(0)
• 'w+' creates a new file or truncates an existing file, then opens it for reading and writing.

with open('some/path.txt', 'r+') as f:
    content = f.read()
    f.write('some other contents')

with open('some/path.txt', 'w+') as f:  # 
    content = f.read()
    f.write('some other contents')

with open('some/path.txt', 'a+') as f:
    content = f.read()
    f.write('some other contents')

That means there is no need to do the following:

# don't do this. use 'w+', 'r+', or 'a+' instead
path = 'some/path.txt'
with open(path) as f_in:
    content = f_in.read()
    with open(path, 'w') as f_out:
        f_out.write('some other contents')