15.2. io — Core tools for working with streams

Python v2.7.2

15.2. io — Core tools for working with streams

New in version 2.6.

The io module provides the Python interfaces to stream handling. Under Python 2.x, this is proposed as an alternative to the built-in file object, but in Python 3.x it is the default interface to access files and streams.

Note

Since this module has been designed primarily for Python 3.x, you have to be aware that all uses of “bytes” in this document refer to the str type (of which bytes is an alias), and all uses of “text” refer to the unicode type. Furthermore, those two types are not interchangeable in the io APIs.

At the top of the I/O hierarchy is the abstract base class IOBase. It defines the basic interface to a stream. Note, however, that there is no separation between reading and writing to streams; implementations are allowed to raise an IOError if they do not support a given operation.

Extending IOBase is RawIOBase which deals simply with the reading and writing of raw bytes to a stream. FileIO subclasses RawIOBase to provide an interface to files in the machine’s file system.

BufferedIOBase deals with buffering on a raw byte stream (RawIOBase). Its subclasses, BufferedWriter, BufferedReader, and BufferedRWPair buffer streams that are readable, writable, and both readable and writable. BufferedRandom provides a buffered interface to random access streams. BytesIO is a simple stream of in-memory bytes.

Another IOBase subclass, TextIOBase, deals with streams whose bytes represent text, and handles encoding and decoding from and to unicode strings. TextIOWrapper, which extends it, is a buffered text interface to a buffered raw stream (BufferedIOBase). Finally, StringIO is an in-memory stream for unicode text.

Argument names are not part of the specification, and only the arguments of open() are intended to be used as keyword arguments.

15.2.1. Module Interface

io.DEFAULT_BUFFER_SIZE
An int containing the default buffer size used by the module’s buffered I/O classes. open() uses the file’s blksize (as obtained by os.stat()) if possible.
io.open(file, mode='r', buffering=-1, encoding=None, errors=None, newline=None, closefd=True)

Open file and return a corresponding stream. If the file cannot be opened, an IOError is raised.

file is either a string giving the pathname (absolute or relative to the current working directory) of the file to be opened or an integer file descriptor of the file to be wrapped. (If a file descriptor is given, it is closed when the returned I/O object is closed, unless closefd is set to False.)

mode is an optional string that specifies the mode in which the file is opened. It defaults to 'r' which means open for reading in text mode. Other common values are 'w' for writing (truncating the file if it already exists), and 'a' for appending (which on some Unix systems, means that all writes append to the end of the file regardless of the current seek position). In text mode, if encoding is not specified the encoding used is platform dependent. (For reading and writing raw bytes use binary mode and leave encoding unspecified.) The available modes are:

Character Meaning
'r' open for reading (default)
'w' open for writing, truncating the file first
'a' open for writing, appending to the end of the file if it exists
'b' binary mode
't' text mode (default)
'+' open a disk file for updating (reading and writing)
'U' universal newline mode (for backwards compatibility; should not be used in new code)

The default mode is 'rt' (open for reading text). For binary random access, the mode 'w+b' opens and truncates the file to 0 bytes, while 'r+b' opens the file without truncation.

Python distinguishes between files opened in binary and text modes, even when the underlying operating system doesn’t. Files opened in binary mode (including 'b' in the mode argument) return contents as bytes objects without any decoding. In text mode (the default, or when 't' is included in the mode argument), the contents of the file are returned as unicode strings, the bytes having been first decoded using a platform-dependent encoding or using the specified encoding if given.

buffering is an optional integer used to set the buffering policy. Pass 0 to switch buffering off (only allowed in binary mode), 1 to select line buffering (only usable in text mode), and an integer > 1 to indicate the size of a fixed-size chunk buffer. When no buffering argument is given, the default buffering policy works as follows:

  • Binary files are buffered in fixed-size chunks; the size of the buffer is chosen using a heuristic trying to determine the underlying device’s “block size” and falling back on DEFAULT_BUFFER_SIZE. On many systems, the buffer will typically be 4096 or 8192 bytes long.
  • “Interactive” text files (files for which isatty() returns True) use line buffering. Other text files use the policy described above for binary files.

encoding is the name of the encoding used to decode or encode the file. This should only be used in text mode. The default encoding is platform dependent (whatever locale.getpreferredencoding() returns), but any encoding supported by Python can be used. See the codecs module for the list of supported encodings.

errors is an optional string that specifies how encoding and decoding errors are to be handled–this cannot be used in binary mode. Pass 'strict' to raise a ValueError exception if there is an encoding error (the default of None has the same effect), or pass 'ignore' to ignore errors. (Note that ignoring encoding errors can lead to data loss.) 'replace' causes a replacement marker (such as '?') to be inserted where there is malformed data. When writing, 'xmlcharrefreplace' (replace with the appropriate XML character reference) or 'backslashreplace' (replace with backslashed escape sequences) can be used. Any other error handling name that has been registered with codecs.register_error() is also valid.

newline controls how universal newlines works (it only applies to text mode). It can be None, '', '\n', '\r', and '\r\n'. It works as follows:

  • On input, if newline is None, universal newlines mode is enabled. Lines in the input can end in '\n', '\r', or '\r\n', and these are translated into '\n' before being returned to the caller. If it is '', universal newline mode is enabled, but line endings are returned to the caller untranslated. If it has any of the other legal values, input lines are only terminated by the given string, and the line ending is returned to the caller untranslated.
  • On output, if newline is None, any '\n' characters written are translated to the system default line separator, os.linesep. If newline is '', no translation takes place. If newline is any of the other legal values, any '\n' characters written are translated to the given string.

If closefd is False and a file descriptor rather than a filename was given, the underlying file descriptor will be kept open when the file is closed. If a filename is given closefd has no effect and must be True (the default).

The type of file object returned by the open() function depends on the mode. When open() is used to open a file in a text mode ('w', 'r', 'wt', 'rt', etc.), it returns a subclass of TextIOBase (specifically TextIOWrapper). When used to open a file in a binary mode with buffering, the returned class is a subclass of BufferedIOBase. The exact class varies: in read binary mode, it returns a BufferedReader; in write binary and append binary modes, it returns a BufferedWriter, and in read/write mode, it returns a BufferedRandom. When buffering is disabled, the raw stream, a subclass of RawIOBase, FileIO, is returned.

It is also possible to use an unicode or bytes string as a file for both reading and writing. For unicode strings StringIO can be used like a file opened in text mode, and for bytes a BytesIO can be used like a file opened in a binary mode.

exception io.BlockingIOError

Error raised when blocking would occur on a non-blocking stream. It inherits IOError.

In addition to those of IOError, BlockingIOError has one attribute:

characters_written
An integer containing the number of characters written to the stream before it blocked.
exception io.UnsupportedOperation
An exception inheriting IOError and ValueError that is raised when an unsupported operation is called on a stream.

15.2.2. I/O Base Classes

class io.IOBase

The abstract base class for all I/O classes, acting on streams of bytes. There is no public constructor.

This class provides empty abstract implementations for many methods that derived classes can override selectively; the default implementations represent a file that cannot be read, written or seeked.

Even though IOBase does not declare read(), readinto(), or write() because their signatures will vary, implementations and clients should consider those methods part of the interface. Also, implementations may raise a IOError when operations they do not support are called.

The basic type used for binary data read from or written to a file is bytes (also known as str). bytearrays are accepted too, and in some cases (such as readinto) required. Text I/O classes work with unicode data.

Note that calling any method (even inquiries) on a closed stream is undefined. Implementations may raise IOError in this case.

IOBase (and its subclasses) support the iterator protocol, meaning that an IOBase object can be iterated over yielding the lines in a stream. Lines are defined slightly differently depending on whether the stream is a binary stream (yielding bytes), or a text stream (yielding unicode strings). See readline() below.

IOBase is also a context manager and therefore supports the with statement. In this example, file is closed after the with statement’s suite is finished—even if an exception occurs:

with io.open('spam.txt', 'w') as file:
    file.write(u'Spam and eggs!')

IOBase provides these data attributes and methods:

close()

Flush and close this stream. This method has no effect if the file is already closed. Once the file is closed, any operation on the file (e.g. reading or writing) will raise a ValueError.

As a convenience, it is allowed to call this method more than once; only the first call, however, will have an effect.

closed
True if the stream is closed.
fileno()
Return the underlying file descriptor (an integer) of the stream if it exists. An IOError is raised if the IO object does not use a file descriptor.
flush()
Flush the write buffers of the stream if applicable. This does nothing for read-only and non-blocking streams.
isatty()
Return True if the stream is interactive (i.e., connected to a terminal/tty device).
readable()
Return True if the stream can be read from. If False, read() will raise IOError.
readline(limit=-1)

Read and return one line from the stream. If limit is specified, at most limit bytes will be read.

The line terminator is always b'\n' for binary files; for text files, the newlines argument to open() can be used to select the line terminator(s) recognized.

readlines(hint=-1)
Read and return a list of lines from the stream. hint can be specified to control the number of lines read: no more lines will be read if the total size (in bytes/characters) of all lines so far exceeds hint.
seek(offset, whence=SEEK_SET)

Change the stream position to the given byte offset. offset is interpreted relative to the position indicated by whence. Values for whence are:

  • SEEK_SET or 0 – start of the stream (the default); offset should be zero or positive
  • SEEK_CUR or 1 – current stream position; offset may be negative
  • SEEK_END or 2 – end of the stream; offset is usually negative

Return the new absolute position.

New in version 2.7: The SEEK_* constants

seekable()
Return True if the stream supports random access. If False, seek(), tell() and truncate() will raise IOError.
tell()
Return the current stream position.
truncate(size=None)
Resize the stream to the given size in bytes (or the current position if size is not specified). The current stream position isn’t changed. This resizing can extend or reduce the current file size. In case of extension, the contents of the new file area depend on the platform (on most systems, additional bytes are zero-filled, on Windows they’re undetermined). The new file size is returned.
writable()
Return True if the stream supports writing. If False, write() and truncate() will raise IOError.
writelines(lines)
Write a list of lines to the stream. Line separators are not added, so it is usual for each of the lines provided to have a line separator at the end.
class io.RawIOBase

Base class for raw binary I/O. It inherits IOBase. There is no public constructor.

Raw binary I/O typically provides low-level access to an underlying OS device or API, and does not try to encapsulate it in high-level primitives (this is left to Buffered I/O and Text I/O, described later in this page).

In addition to the attributes and methods from IOBase, RawIOBase provides the following methods:

read(n=-1)

Read up to n bytes from the object and return them. As a convenience, if n is unspecified or -1, readall() is called. Otherwise, only one system call is ever made. Fewer than n bytes may be returned if the operating system call returns fewer than n bytes.

If 0 bytes are returned, and n was not 0, this indicates end of file. If the object is in non-blocking mode and no bytes are available, None is returned.

readall()
Read and return all the bytes from the stream until EOF, using multiple calls to the stream if necessary.
readinto(b)
Read up to len(b) bytes into bytearray b and return the number of bytes read. If the object is in non-blocking mode and no bytes are available, None is returned.
write(b)
Write the given bytes or bytearray object, b, to the underlying raw stream and return the number of bytes written. This can be less than len(b), depending on specifics of the underlying raw stream, and especially if it is in non-blocking mode. None is returned if the raw stream is set not to block and no single byte could be readily written to it.
class io.BufferedIOBase

Base class for binary streams that support some kind of buffering. It inherits IOBase. There is no public constructor.

The main difference with RawIOBase is that methods read(), readinto() and write() will try (respectively) to read as much input as requested or to consume all given output, at the expense of making perhaps more than one system call.

In addition, those methods can raise BlockingIOError if the underlying raw stream is in non-blocking mode and cannot take or give enough data; unlike their RawIOBase counterparts, they will never return None.

Besides, the read() method does not have a default implementation that defers to readinto().

A typical BufferedIOBase implementation should not inherit from a RawIOBase implementation, but wrap one, like BufferedWriter and BufferedReader do.

BufferedIOBase provides or overrides these members in addition to those from IOBase:

raw
The underlying raw stream (a RawIOBase instance) that BufferedIOBase deals with. This is not part of the BufferedIOBase API and may not exist on some implementations.
detach()

Separate the underlying raw stream from the buffer and return it.

After the raw stream has been detached, the buffer is in an unusable state.

Some buffers, like BytesIO, do not have the concept of a single raw stream to return from this method. They raise UnsupportedOperation.

New in version 2.7.

read(n=-1)

Read and return up to n bytes. If the argument is omitted, None, or negative, data is read and returned until EOF is reached. An empty bytes object is returned if the stream is already at EOF.

If the argument is positive, and the underlying raw stream is not interactive, multiple raw reads may be issued to satisfy the byte count (unless EOF is reached first). But for interactive raw streams, at most one raw read will be issued, and a short result does not imply that EOF is imminent.

A BlockingIOError is raised if the underlying raw stream is in non blocking-mode, and has no data available at the moment.

read1(n=-1)
Read and return up to n bytes, with at most one call to the underlying raw stream’s read() method. This can be useful if you are implementing your own buffering on top of a BufferedIOBase object.
readinto(b)

Read up to len(b) bytes into bytearray b and return the number of bytes read.

Like read(), multiple reads may be issued to the underlying raw stream, unless the latter is ‘interactive’.

A BlockingIOError is raised if the underlying raw stream is in non blocking-mode, and has no data available at the moment.

write(b)

Write the given bytes or bytearray object, b and return the number of bytes written (never less than len(b), since if the write fails an IOError will be raised). Depending on the actual implementation, these bytes may be readily written to the underlying stream, or held in a buffer for performance and latency reasons.

When in non-blocking mode, a BlockingIOError is raised if the data needed to be written to the raw stream but it couldn’t accept all the data without blocking.

15.2.3. Raw File I/O

class io.FileIO(name, mode='r', closefd=True)

FileIO represents an OS-level file containing bytes data. It implements the RawIOBase interface (and therefore the IOBase interface, too).

The name can be one of two things:

  • a string representing the path to the file which will be opened;
  • an integer representing the number of an existing OS-level file descriptor to which the resulting FileIO object will give access.

The mode can be 'r', 'w' or 'a' for reading (default), writing, or appending. The file will be created if it doesn’t exist when opened for writing or appending; it will be truncated when opened for writing. Add a '+' to the mode to allow simultaneous reading and writing.

The read() (when called with a positive argument), readinto() and write() methods on this class will only make one system call.

In addition to the attributes and methods from IOBase and RawIOBase, FileIO provides the following data attributes and methods:

mode
The mode as given in the constructor.
name
The file name. This is the file descriptor of the file when no name is given in the constructor.

15.2.4. Buffered Streams

Buffered I/O streams provide a higher-level interface to an I/O device than raw I/O does.

class io.BytesIO([initial_bytes])

A stream implementation using an in-memory bytes buffer. It inherits BufferedIOBase.

The argument initial_bytes is an optional initial bytes.

BytesIO provides or overrides these methods in addition to those from BufferedIOBase and IOBase:

getvalue()
Return bytes containing the entire contents of the buffer.
read1()
In BytesIO, this is the same as read().
class io.BufferedReader(raw, buffer_size=DEFAULT_BUFFER_SIZE)

A buffer providing higher-level access to a readable, sequential RawIOBase object. It inherits BufferedIOBase. When reading data from this object, a larger amount of data may be requested from the underlying raw stream, and kept in an internal buffer. The buffered data can then be returned directly on subsequent reads.

The constructor creates a BufferedReader for the given readable raw stream and buffer_size. If buffer_size is omitted, DEFAULT_BUFFER_SIZE is used.

BufferedReader provides or overrides these methods in addition to those from BufferedIOBase and IOBase:

peek([n])
Return bytes from the stream without advancing the position. At most one single read on the raw stream is done to satisfy the call. The number of bytes returned may be less or more than requested.
read([n])
Read and return n bytes, or if n is not given or negative, until EOF or if the read call would block in non-blocking mode.
read1(n)
Read and return up to n bytes with only one call on the raw stream. If at least one byte is buffered, only buffered bytes are returned. Otherwise, one raw stream read call is made.
class io.BufferedWriter(raw, buffer_size=DEFAULT_BUFFER_SIZE)

A buffer providing higher-level access to a writeable, sequential RawIOBase object. It inherits BufferedIOBase. When writing to this object, data is normally held into an internal buffer. The buffer will be written out to the underlying RawIOBase object under various conditions, including:

  • when the buffer gets too small for all pending data;
  • when flush() is called;
  • when a seek() is requested (for BufferedRandom objects);
  • when the BufferedWriter object is closed or destroyed.

The constructor creates a BufferedWriter for the given writeable raw stream. If the buffer_size is not given, it defaults to DEFAULT_BUFFER_SIZE.

A third argument, max_buffer_size, is supported, but unused and deprecated.

BufferedWriter provides or overrides these methods in addition to those from BufferedIOBase and IOBase:

flush()
Force bytes held in the buffer into the raw stream. A BlockingIOError should be raised if the raw stream blocks.
write(b)
Write the bytes or bytearray object, b and return the number of bytes written. When in non-blocking mode, a BlockingIOError is raised if the buffer needs to be written out but the raw stream blocks.
class io.BufferedRWPair(reader, writer, buffer_size=DEFAULT_BUFFER_SIZE)

A buffered I/O object giving a combined, higher-level access to two sequential RawIOBase objects: one readable, the other writeable. It is useful for pairs of unidirectional communication channels (pipes, for instance). It inherits BufferedIOBase.

reader and writer are RawIOBase objects that are readable and writeable respectively. If the buffer_size is omitted it defaults to DEFAULT_BUFFER_SIZE.

A fourth argument, max_buffer_size, is supported, but unused and deprecated.

BufferedRWPair implements all of BufferedIOBase‘s methods except for detach(), which raises UnsupportedOperation.

class io.BufferedRandom(raw, buffer_size=DEFAULT_BUFFER_SIZE)

A buffered interface to random access streams. It inherits BufferedReader and BufferedWriter, and further supports seek() and tell() functionality.

The constructor creates a reader and writer for a seekable raw stream, given in the first argument. If the buffer_size is omitted it defaults to DEFAULT_BUFFER_SIZE.

A third argument, max_buffer_size, is supported, but unused and deprecated.

BufferedRandom is capable of anything BufferedReader or BufferedWriter can do.

15.2.5. Text I/O

class io.TextIOBase

Base class for text streams. This class provides an unicode character and line based interface to stream I/O. There is no readinto() method because Python’s unicode strings are immutable. It inherits IOBase. There is no public constructor.

TextIOBase provides or overrides these data attributes and methods in addition to those from IOBase:

encoding
The name of the encoding used to decode the stream’s bytes into strings, and to encode strings into bytes.
errors
The error setting of the decoder or encoder.
newlines
A string, a tuple of strings, or None, indicating the newlines translated so far. Depending on the implementation and the initial constructor flags, this may not be available.
buffer
The underlying binary buffer (a BufferedIOBase instance) that TextIOBase deals with. This is not part of the TextIOBase API and may not exist on some implementations.
detach()

Separate the underlying binary buffer from the TextIOBase and return it.

After the underlying buffer has been detached, the TextIOBase is in an unusable state.

Some TextIOBase implementations, like StringIO, may not have the concept of an underlying buffer and calling this method will raise UnsupportedOperation.

New in version 2.7.

read(n)
Read and return at most n characters from the stream as a single unicode. If n is negative or None, reads until EOF.
readline()
Read until newline or EOF and return a single unicode. If the stream is already at EOF, an empty string is returned.
write(s)
Write the unicode string s to the stream and return the number of characters written.
class io.TextIOWrapper(buffer, encoding=None, errors=None, newline=None, line_buffering=False)

A buffered text stream over a BufferedIOBase binary stream. It inherits TextIOBase.

encoding gives the name of the encoding that the stream will be decoded or encoded with. It defaults to locale.getpreferredencoding().

errors is an optional string that specifies how encoding and decoding errors are to be handled. Pass 'strict' to raise a ValueError exception if there is an encoding error (the default of None has the same effect), or pass 'ignore' to ignore errors. (Note that ignoring encoding errors can lead to data loss.) 'replace' causes a replacement marker (such as '?') to be inserted where there is malformed data. When writing, 'xmlcharrefreplace' (replace with the appropriate XML character reference) or 'backslashreplace' (replace with backslashed escape sequences) can be used. Any other error handling name that has been registered with codecs.register_error() is also valid.

newline can be None, '', '\n', '\r', or '\r\n'. It controls the handling of line endings. If it is None, universal newlines is enabled. With this enabled, on input, the lines endings '\n', '\r', or '\r\n' are translated to '\n' before being returned to the caller. Conversely, on output, '\n' is translated to the system default line separator, os.linesep. If newline is any other of its legal values, that newline becomes the newline when the file is read and it is returned untranslated. On output, '\n' is converted to the newline.

If line_buffering is True, flush() is implied when a call to write contains a newline character.

TextIOWrapper provides one attribute in addition to those of TextIOBase and its parents:

line_buffering
Whether line buffering is enabled.
class io.StringIO(initial_value=u'', newline=None)

An in-memory stream for unicode text. It inherits TextIOWrapper.

The initial value of the buffer (an empty unicode string by default) can be set by providing initial_value. The newline argument works like that of TextIOWrapper. The default is to do no newline translation.

StringIO provides this method in addition to those from TextIOWrapper and its parents:

getvalue()
Return a unicode containing the entire contents of the buffer at any time before the StringIO object’s close() method is called.

Example usage:

import io

output = io.StringIO()
output.write(u'First line.\n')
output.write(u'Second line.\n')

# Retrieve file contents -- this will be
# u'First line.\nSecond line.\n'
contents = output.getvalue()

# Close object and discard memory buffer --
# .getvalue() will now raise an exception.
output.close()
class io.IncrementalNewlineDecoder
A helper codec that decodes newlines for universal newlines mode. It inherits codecs.IncrementalDecoder.

15.2.6. Advanced topics

Here we will discuss several advanced topics pertaining to the concrete I/O implementations described above.

15.2.6.1. Performance

15.2.6.1.1. Binary I/O

By reading and writing only large chunks of data even when the user asks for a single byte, buffered I/O is designed to hide any inefficiency in calling and executing the operating system’s unbuffered I/O routines. The gain will vary very much depending on the OS and the kind of I/O which is performed (for example, on some contemporary OSes such as Linux, unbuffered disk I/O can be as fast as buffered I/O). The bottom line, however, is that buffered I/O will offer you predictable performance regardless of the platform and the backing device. Therefore, it is most always preferable to use buffered I/O rather than unbuffered I/O.

15.2.6.1.2. Text I/O

Text I/O over a binary storage (such as a file) is significantly slower than binary I/O over the same storage, because it implies conversions from unicode to binary data using a character codec. This can become noticeable if you handle huge amounts of text data (for example very large log files). Also, TextIOWrapper.tell() and TextIOWrapper.seek() are both quite slow due to the reconstruction algorithm used.

StringIO, however, is a native in-memory unicode container and will exhibit similar speed to BytesIO.

15.2.6.2. Multi-threading

FileIO objects are thread-safe to the extent that the operating system calls (such as read(2) under Unix) they are wrapping are thread-safe too.

Binary buffered objects (instances of BufferedReader, BufferedWriter, BufferedRandom and BufferedRWPair) protect their internal structures using a lock; it is therefore safe to call them from multiple threads at once.

TextIOWrapper objects are not thread-safe.

15.2.6.3. Reentrancy

Binary buffered objects (instances of BufferedReader, BufferedWriter, BufferedRandom and BufferedRWPair) are not reentrant. While reentrant calls will not happen in normal situations, they can arise if you are doing I/O in a signal handler. If it is attempted to enter a buffered object again while already being accessed from the same thread, then a RuntimeError is raised.

The above implicitly extends to text files, since the open() function will wrap a buffered object inside a TextIOWrapper. This includes standard streams and therefore affects the built-in function print() as well.