tarfile --- Read and write tar archive files


The tarfile module makes it possible to read and write tar archives, including those using gzip, bz2 and lzma compression. Use the zipfile module to read or write .zip files, or the higher-level functions in shutil.

Some facts and figures:

  • reads and writes gzip, bz2 and lzma compressed archives if the respective modules are available.

  • read/write support for the POSIX.1-1988 (ustar) format.

  • read/write support for the GNU tar format including longname and longlink extensions, read-only support for all variants of the sparse extension including restoration of sparse files.

  • read/write support for the POSIX.1-2001 (pax) format.

  • handles directories, regular files, hardlinks, symbolic links, fifos, character devices and block devices and is able to acquire and restore file information like timestamp, access permissions and owner.

在 3.3 版的變更: Added support for lzma compression.

tarfile.open(name=None, mode='r', fileobj=None, bufsize=10240, **kwargs)

Return a TarFile object for the pathname name. For detailed information on TarFile objects and the keyword arguments that are allowed, see TarFile 物件.

mode has to be a string of the form 'filemode[:compression]', it defaults to 'r'. Here is a full list of mode combinations:



'r' 'r:*'

Open for reading with transparent compression (recommended).


Open for reading exclusively without compression.


Open for reading with gzip compression.


Open for reading with bzip2 compression.


Open for reading with lzma compression.


Create a tarfile exclusively without compression. Raise a FileExistsError exception if it already exists.


Create a tarfile with gzip compression. Raise a FileExistsError exception if it already exists.


Create a tarfile with bzip2 compression. Raise a FileExistsError exception if it already exists.


Create a tarfile with lzma compression. Raise a FileExistsError exception if it already exists.

'a' 'a:'

Open for appending with no compression. The file is created if it does not exist.

'w' 'w:'

Open for uncompressed writing.


Open for gzip compressed writing.


Open for bzip2 compressed writing.


Open for lzma compressed writing.

Note that 'a:gz', 'a:bz2' or 'a:xz' is not possible. If mode is not suitable to open a certain (compressed) file for reading, ReadError is raised. Use mode 'r' to avoid this. If a compression method is not supported, CompressionError is raised.

If fileobj is specified, it is used as an alternative to a file object opened in binary mode for name. It is supposed to be at position 0.

For modes 'w:gz', 'r:gz', 'w:bz2', 'r:bz2', 'x:gz', 'x:bz2', tarfile.open() accepts the keyword argument compresslevel (default 9) to specify the compression level of the file.

For modes 'w:xz' and 'x:xz', tarfile.open() accepts the keyword argument preset to specify the compression level of the file.

For special purposes, there is a second format for mode: 'filemode|[compression]'. tarfile.open() will return a TarFile object that processes its data as a stream of blocks. No random seeking will be done on the file. If given, fileobj may be any object that has a read() or write() method (depending on the mode) that works with bytes. bufsize specifies the blocksize and defaults to 20 * 512 bytes. Use this variant in combination with e.g. sys.stdin.buffer, a socket file object or a tape device. However, such a TarFile object is limited in that it does not allow random access, see 範例. The currently possible modes:




Open a stream of tar blocks for reading with transparent compression.


Open a stream of uncompressed tar blocks for reading.


Open a gzip compressed stream for reading.


Open a bzip2 compressed stream for reading.


Open an lzma compressed stream for reading.


Open an uncompressed stream for writing.


Open a gzip compressed stream for writing.


Open a bzip2 compressed stream for writing.


Open an lzma compressed stream for writing.

在 3.5 版的變更: The 'x' (exclusive creation) mode was added.

在 3.6 版的變更: The name parameter accepts a path-like object.

class tarfile.TarFile

Class for reading and writing tar archives. Do not use this class directly: use tarfile.open() instead. See TarFile 物件.


Return True if name is a tar archive file, that the tarfile module can read. name may be a str, file, or file-like object.

在 3.9 版的變更: Support for file and file-like objects.

The tarfile module defines the following exceptions:

exception tarfile.TarError

Base class for all tarfile exceptions.

exception tarfile.ReadError

Is raised when a tar archive is opened, that either cannot be handled by the tarfile module or is somehow invalid.

exception tarfile.CompressionError

Is raised when a compression method is not supported or when the data cannot be decoded properly.

exception tarfile.StreamError

Is raised for the limitations that are typical for stream-like TarFile objects.

exception tarfile.ExtractError

Is raised for non-fatal errors when using TarFile.extract(), but only if TarFile.errorlevel== 2.

exception tarfile.HeaderError

Is raised by TarInfo.frombuf() if the buffer it gets is invalid.

exception tarfile.FilterError

Base class for members refused by filters.


Information about the member that the filter refused to extract, as TarInfo.

exception tarfile.AbsolutePathError

Raised to refuse extracting a member with an absolute path.

exception tarfile.OutsideDestinationError

Raised to refuse extracting a member outside the destination directory.

exception tarfile.SpecialFileError

Raised to refuse extracting a special file (e.g. a device or pipe).

exception tarfile.AbsoluteLinkError

Raised to refuse extracting a symbolic link with an absolute path.

exception tarfile.LinkOutsideDestinationError

Raised to refuse extracting a symbolic link pointing outside the destination directory.

The following constants are available at the module level:


The default character encoding: 'utf-8' on Windows, the value returned by sys.getfilesystemencoding() otherwise.


A regular file type.


A link (inside tarfile) type.


A symbolic link type.


A character special device type.


A block special device type.


A directory type.


A FIFO special device type.


A contiguous file type.


A GNU tar longname type.

A GNU tar longlink type.


A GNU tar sparse file type.

Each of the following constants defines a tar archive format that the tarfile module is able to create. See section Supported tar formats for details.


POSIX.1-1988 (ustar) format.


GNU tar format.


POSIX.1-2001 (pax) format.


The default format for creating archives. This is currently PAX_FORMAT.

在 3.8 版的變更: The default format for new archives was changed to PAX_FORMAT from GNU_FORMAT.


zipfile 模組

Documentation of the zipfile standard module.

Archiving operations

Documentation of the higher-level archiving facilities provided by the standard shutil module.

GNU tar manual, Basic Tar Format

Documentation for tar archive files, including GNU tar extensions.

TarFile 物件

The TarFile object provides an interface to a tar archive. A tar archive is a sequence of blocks. An archive member (a stored file) is made up of a header block followed by data blocks. It is possible to store a file in a tar archive several times. Each archive member is represented by a TarInfo object, see TarInfo 物件 for details.

A TarFile object can be used as a context manager in a with statement. It will automatically be closed when the block is completed. Please note that in the event of an exception an archive opened for writing will not be finalized; only the internally used file object will be closed. See the 範例 section for a use case.

在 3.2 版新加入: Added support for the context management protocol.

class tarfile.TarFile(name=None, mode='r', fileobj=None, format=DEFAULT_FORMAT, tarinfo=TarInfo, dereference=False, ignore_zeros=False, encoding=ENCODING, errors='surrogateescape', pax_headers=None, debug=0, errorlevel=1)

All following arguments are optional and can be accessed as instance attributes as well.

name is the pathname of the archive. name may be a path-like object. It can be omitted if fileobj is given. In this case, the file object's name attribute is used if it exists.

mode is either 'r' to read from an existing archive, 'a' to append data to an existing file, 'w' to create a new file overwriting an existing one, or 'x' to create a new file only if it does not already exist.

If fileobj is given, it is used for reading or writing data. If it can be determined, mode is overridden by fileobj's mode. fileobj will be used from position 0.


fileobj is not closed, when TarFile is closed.

format controls the archive format for writing. It must be one of the constants USTAR_FORMAT, GNU_FORMAT or PAX_FORMAT that are defined at module level. When reading, format will be automatically detected, even if different formats are present in a single archive.

The tarinfo argument can be used to replace the default TarInfo class with a different one.

If dereference is False, add symbolic and hard links to the archive. If it is True, add the content of the target files to the archive. This has no effect on systems that do not support symbolic links.

If ignore_zeros is False, treat an empty block as the end of the archive. If it is True, skip empty (and invalid) blocks and try to get as many members as possible. This is only useful for reading concatenated or damaged archives.

debug can be set from 0 (no debug messages) up to 3 (all debug messages). The messages are written to sys.stderr.

errorlevel controls how extraction errors are handled, see the corresponding attribute.

The encoding and errors arguments define the character encoding to be used for reading or writing the archive and how conversion errors are going to be handled. The default settings will work for most users. See section Unicode issues for in-depth information.

The pax_headers argument is an optional dictionary of strings which will be added as a pax global header if format is PAX_FORMAT.

在 3.2 版的變更: Use 'surrogateescape' as the default for the errors argument.

在 3.5 版的變更: The 'x' (exclusive creation) mode was added.

在 3.6 版的變更: The name parameter accepts a path-like object.

classmethod TarFile.open(...)

Alternative constructor. The tarfile.open() function is actually a shortcut to this classmethod.


Return a TarInfo object for member name. If name can not be found in the archive, KeyError is raised.


If a member occurs more than once in the archive, its last occurrence is assumed to be the most up-to-date version.


Return the members of the archive as a list of TarInfo objects. The list has the same order as the members in the archive.


Return the members as a list of their names. It has the same order as the list returned by getmembers().

TarFile.list(verbose=True, *, members=None)

Print a table of contents to sys.stdout. If verbose is False, only the names of the members are printed. If it is True, output similar to that of ls -l is produced. If optional members is given, it must be a subset of the list returned by getmembers().

在 3.5 版的變更: 新增 members 參數。


Return the next member of the archive as a TarInfo object, when TarFile is opened for reading. Return None if there is no more available.

TarFile.extractall(path='.', members=None, *, numeric_owner=False, filter=None)

Extract all members from the archive to the current working directory or directory path. If optional members is given, it must be a subset of the list returned by getmembers(). Directory information like owner, modification time and permissions are set after all members have been extracted. This is done to work around two problems: A directory's modification time is reset each time a file is created in it. And, if a directory's permissions do not allow writing, extracting files to it will fail.

If numeric_owner is True, the uid and gid numbers from the tarfile are used to set the owner/group for the extracted files. Otherwise, the named values from the tarfile are used.

The filter argument, which was added in Python 3.11.4, specifies how members are modified or rejected before extraction. See Extraction filters for details. It is recommended to set this explicitly depending on which tar features you need to support.


Never extract archives from untrusted sources without prior inspection. It is possible that files are created outside of path, e.g. members that have absolute filenames starting with "/" or filenames with two dots "..".

Set filter='data' to prevent the most dangerous security issues, and read the Extraction filters section for details.

在 3.5 版的變更: 新增 numeric_owner 參數。

在 3.6 版的變更: The path parameter accepts a path-like object.

在 3.11.4 版的變更: 新增 filter 參數。

TarFile.extract(member, path='', set_attrs=True, *, numeric_owner=False, filter=None)

Extract a member from the archive to the current working directory, using its full name. Its file information is extracted as accurately as possible. member may be a filename or a TarInfo object. You can specify a different directory using path. path may be a path-like object. File attributes (owner, mtime, mode) are set unless set_attrs is false.

The numeric_owner and filter arguments are the same as for extractall().


The extract() method does not take care of several extraction issues. In most cases you should consider using the extractall() method.


參閱 extractall() 的警告。

Set filter='data' to prevent the most dangerous security issues, and read the Extraction filters section for details.

在 3.2 版的變更: 增加 set_attrs 參數。

在 3.5 版的變更: 新增 numeric_owner 參數。

在 3.6 版的變更: The path parameter accepts a path-like object.

在 3.11.4 版的變更: 新增 filter 參數。


Extract a member from the archive as a file object. member may be a filename or a TarInfo object. If member is a regular file or a link, an io.BufferedReader object is returned. For all other existing members, None is returned. If member does not appear in the archive, KeyError is raised.

在 3.3 版的變更: Return an io.BufferedReader object.

TarFile.errorlevel: int

If errorlevel is 0, errors are ignored when using TarFile.extract() and TarFile.extractall(). Nevertheless, they appear as error messages in the debug output when debug is greater than 0. If 1 (the default), all fatal errors are raised as OSError or FilterError exceptions. If 2, all non-fatal errors are raised as TarError exceptions as well.

Some exceptions, e.g. ones caused by wrong argument types or data corruption, are always raised.

Custom extraction filters should raise FilterError for fatal errors and ExtractError for non-fatal ones.

Note that when an exception is raised, the archive may be partially extracted. It is the user’s responsibility to clean up.


在 3.11.4 版新加入.

The extraction filter used as a default for the filter argument of extract() and extractall().

The attribute may be None or a callable. String names are not allowed for this attribute, unlike the filter argument to extract().

If extraction_filter is None (the default), calling an extraction method without a filter argument will use the fully_trusted filter for compatibility with previous Python versions.

In Python 3.12+, leaving extraction_filter=None will emit a DeprecationWarning.

In Python 3.14+, leaving extraction_filter=None will cause extraction methods to use the data filter by default.

The attribute may be set on instances or overridden in subclasses. It also is possible to set it on the TarFile class itself to set a global default, although, since it affects all uses of tarfile, it is best practice to only do so in top-level applications or site configuration. To set a global default this way, a filter function needs to be wrapped in staticmethod() to prevent injection of a self argument.

TarFile.add(name, arcname=None, recursive=True, *, filter=None)

Add the file name to the archive. name may be any type of file (directory, fifo, symbolic link, etc.). If given, arcname specifies an alternative name for the file in the archive. Directories are added recursively by default. This can be avoided by setting recursive to False. Recursion adds entries in sorted order. If filter is given, it should be a function that takes a TarInfo object argument and returns the changed TarInfo object. If it instead returns None the TarInfo object will be excluded from the archive. See 範例 for an example.

在 3.2 版的變更: 新增 filter 參數。

在 3.7 版的變更: Recursion adds entries in sorted order.

TarFile.addfile(tarinfo, fileobj=None)

Add the TarInfo object tarinfo to the archive. If fileobj is given, it should be a binary file, and tarinfo.size bytes are read from it and added to the archive. You can create TarInfo objects directly, or by using gettarinfo().

TarFile.gettarinfo(name=None, arcname=None, fileobj=None)

Create a TarInfo object from the result of os.stat() or equivalent on an existing file. The file is either named by name, or specified as a file object fileobj with a file descriptor. name may be a path-like object. If given, arcname specifies an alternative name for the file in the archive, otherwise, the name is taken from fileobj’s name attribute, or the name argument. The name should be a text string.

You can modify some of the TarInfo’s attributes before you add it using addfile(). If the file object is not an ordinary file object positioned at the beginning of the file, attributes such as size may need modifying. This is the case for objects such as GzipFile. The name may also be modified, in which case arcname could be a dummy string.

在 3.6 版的變更: The name parameter accepts a path-like object.


Close the TarFile. In write mode, two finishing zero blocks are appended to the archive.

TarFile.pax_headers: dict

A dictionary containing key-value pairs of pax global headers.

TarInfo 物件

A TarInfo object represents one member in a TarFile. Aside from storing all required attributes of a file (like file type, size, time, permissions, owner etc.), it provides some useful methods to determine its type. It does not contain the file's data itself.

TarInfo objects are returned by TarFile's methods getmember(), getmembers() and gettarinfo().

Modifying the objects returned by getmember() or getmembers() will affect all subsequent operations on the archive. For cases where this is unwanted, you can use copy.copy() or call the replace() method to create a modified copy in one step.

Several attributes can be set to None to indicate that a piece of metadata is unused or unknown. Different TarInfo methods handle None differently:

在 3.11.4 版的變更: Added replace() and handling of None.

class tarfile.TarInfo(name='')

Create a TarInfo object.

classmethod TarInfo.frombuf(buf, encoding, errors)

Create and return a TarInfo object from string buffer buf.

Raises HeaderError if the buffer is invalid.

classmethod TarInfo.fromtarfile(tarfile)

Read the next member from the TarFile object tarfile and return it as a TarInfo object.

TarInfo.tobuf(format=DEFAULT_FORMAT, encoding=ENCODING, errors='surrogateescape')

Create a string buffer from a TarInfo object. For information on the arguments see the constructor of the TarFile class.

在 3.2 版的變更: Use 'surrogateescape' as the default for the errors argument.

A TarInfo object has the following public data attributes:

TarInfo.name: str

Name of the archive member.

TarInfo.size: int

Size in bytes.

TarInfo.mtime: int | float

Time of last modification in seconds since the epoch, as in os.stat_result.st_mtime.

在 3.11.4 版的變更: Can be set to None for extract() and extractall(), causing extraction to skip applying this attribute.

TarInfo.mode: int

Permission bits, as for os.chmod().

在 3.11.4 版的變更: Can be set to None for extract() and extractall(), causing extraction to skip applying this attribute.


File type. type is usually one of these constants: REGTYPE, AREGTYPE, LNKTYPE, SYMTYPE, DIRTYPE, FIFOTYPE, CONTTYPE, CHRTYPE, BLKTYPE, GNUTYPE_SPARSE. To determine the type of a TarInfo object more conveniently, use the is*() methods below.

TarInfo.linkname: str

Name of the target file name, which is only present in TarInfo objects of type LNKTYPE and SYMTYPE.

For symbolic links (SYMTYPE), the linkname is relative to the directory that contains the link. For hard links (LNKTYPE), the linkname is relative to the root of the archive.

TarInfo.uid: int

User ID of the user who originally stored this member.

在 3.11.4 版的變更: Can be set to None for extract() and extractall(), causing extraction to skip applying this attribute.

TarInfo.gid: int

Group ID of the user who originally stored this member.

在 3.11.4 版的變更: Can be set to None for extract() and extractall(), causing extraction to skip applying this attribute.

TarInfo.uname: str

User name.

在 3.11.4 版的變更: Can be set to None for extract() and extractall(), causing extraction to skip applying this attribute.

TarInfo.gname: str

Group name.

在 3.11.4 版的變更: Can be set to None for extract() and extractall(), causing extraction to skip applying this attribute.

TarInfo.chksum: int

Header checksum.

TarInfo.devmajor: int

Device major number.

TarInfo.devminor: int

Device minor number.

TarInfo.offset: int

The tar header starts here.

TarInfo.offset_data: int

The file's data starts here.


Sparse member information.

TarInfo.pax_headers: dict

A dictionary containing key-value pairs of an associated pax extended header.

TarInfo.replace(name=..., mtime=..., mode=..., linkname=..., uid=..., gid=..., uname=..., gname=..., deep=True)

在 3.11.4 版新加入.

Return a new copy of the TarInfo object with the given attributes changed. For example, to return a TarInfo with the group name set to 'staff', use:

new_tarinfo = old_tarinfo.replace(gname='staff')

By default, a deep copy is made. If deep is false, the copy is shallow, i.e. pax_headers and any custom attributes are shared with the original TarInfo object.

A TarInfo object also provides some convenient query methods:


Return True if the TarInfo object is a regular file.


Same as isfile().


Return True if it is a directory.


Return True if it is a symbolic link.


Return True if it is a hard link.


Return True if it is a character device.


Return True if it is a block device.


Return True if it is a FIFO.


Return True if it is one of character device, block device or FIFO.

Extraction filters

在 3.11.4 版新加入.

The tar format is designed to capture all details of a UNIX-like filesystem, which makes it very powerful. Unfortunately, the features make it easy to create tar files that have unintended -- and possibly malicious -- effects when extracted. For example, extracting a tar file can overwrite arbitrary files in various ways (e.g. by using absolute paths, .. path components, or symlinks that affect later members).

In most cases, the full functionality is not needed. Therefore, tarfile supports extraction filters: a mechanism to limit functionality, and thus mitigate some of the security issues.


PEP 706

Contains further motivation and rationale behind the design.

The filter argument to TarFile.extract() or extractall() can be:

  • the string 'fully_trusted': Honor all metadata as specified in the archive. Should be used if the user trusts the archive completely, or implements their own complex verification.

  • the string 'tar': Honor most tar-specific features (i.e. features of UNIX-like filesystems), but block features that are very likely to be surprising or malicious. See tar_filter() for details.

  • the string 'data': Ignore or block most features specific to UNIX-like filesystems. Intended for extracting cross-platform data archives. See data_filter() for details.

  • None (default): Use TarFile.extraction_filter.

    If that is also None (the default), the 'fully_trusted' filter will be used (for compatibility with earlier versions of Python).

    In Python 3.12, the default will emit a DeprecationWarning.

    In Python 3.14, the 'data' filter will become the default instead. It's possible to switch earlier; see TarFile.extraction_filter.

  • A callable which will be called for each extracted member with a TarInfo describing the member and the destination path to where the archive is extracted (i.e. the same path is used for all members):

    filter(member: TarInfo, path: str, /) -> TarInfo | None

    The callable is called just before each member is extracted, so it can take the current state of the disk into account. It can:

    • return a TarInfo object which will be used instead of the metadata in the archive, or

    • return None, in which case the member will be skipped, or

    • raise an exception to abort the operation or skip the member, depending on errorlevel. Note that when extraction is aborted, extractall() may leave the archive partially extracted. It does not attempt to clean up.

Default named filters

The pre-defined, named filters are available as functions, so they can be reused in custom filters:

tarfile.fully_trusted_filter(member, path)

Return member unchanged.

This implements the 'fully_trusted' filter.

tarfile.tar_filter(member, path)

Implements the 'tar' filter.

  • Strip leading slashes (/ and os.sep) from filenames.

  • Refuse to extract files with absolute paths (in case the name is absolute even after stripping slashes, e.g. C:/foo on Windows). This raises AbsolutePathError.

  • Refuse to extract files whose absolute path (after following symlinks) would end up outside the destination. This raises OutsideDestinationError.

  • Clear high mode bits (setuid, setgid, sticky) and group/other write bits (S_IWGRP | S_IWOTH).

Return the modified TarInfo member.

tarfile.data_filter(member, path)

Implements the 'data' filter. In addition to what tar_filter does:

  • Refuse to extract links (hard or soft) that link to absolute paths, or ones that link outside the destination.

    This raises AbsoluteLinkError or LinkOutsideDestinationError.

    Note that such files are refused even on platforms that do not support symbolic links.

  • Refuse to extract device files (including pipes). This raises SpecialFileError.

  • For regular files, including hard links:

  • For other files (directories), set mode to None, so that extraction methods skip applying permission bits.

  • Set user and group info (uid, gid, uname, gname) to None, so that extraction methods skip setting it.

Return the modified TarInfo member.

Filter errors

When a filter refuses to extract a file, it will raise an appropriate exception, a subclass of FilterError. This will abort the extraction if TarFile.errorlevel is 1 or more. With errorlevel=0 the error will be logged and the member will be skipped, but extraction will continue.

Hints for further verification

Even with filter='data', tarfile is not suited for extracting untrusted files without prior inspection. Among other issues, the pre-defined filters do not prevent denial-of-service attacks. Users should do additional checks.

Here is an incomplete list of things to consider:

  • Extract to a new temporary directory to prevent e.g. exploiting pre-existing links, and to make it easier to clean up after a failed extraction.

  • When working with untrusted data, use external (e.g. OS-level) limits on disk, memory and CPU usage.

  • Check filenames against an allow-list of characters (to filter out control characters, confusables, foreign path separators, etc.).

  • Check that filenames have expected extensions (discouraging files that execute when you “click on them”, or extension-less files like Windows special device names).

  • Limit the number of extracted files, total size of extracted data, filename length (including symlink length), and size of individual files.

  • Check for files that would be shadowed on case-insensitive filesystems.

Also note that:

  • Tar files may contain multiple versions of the same file. Later ones are expected to overwrite any earlier ones. This feature is crucial to allow updating tape archives, but can be abused maliciously.

  • tarfile does not protect against issues with “live” data, e.g. an attacker tinkering with the destination (or source) directory while extraction (or archiving) is in progress.

Supporting older Python versions

Extraction filters were added to Python 3.12, and are backported to older versions as security updates. To check whether the feature is available, use e.g. hasattr(tarfile, 'data_filter') rather than checking the Python version.

The following examples show how to support Python versions with and without the feature. Note that setting extraction_filter will affect any subsequent operations.

  • Fully trusted archive:

    my_tarfile.extraction_filter = (lambda member, path: member)
  • Use the 'data' filter if available, but revert to Python 3.11 behavior ('fully_trusted') if this feature is not available:

    my_tarfile.extraction_filter = getattr(tarfile, 'data_filter',
                                           (lambda member, path: member))
  • Use the 'data' filter; fail if it is not available:



    my_tarfile.extraction_filter = tarfile.data_filter
  • Use the 'data' filter; warn if it is not available:

    if hasattr(tarfile, 'data_filter'):
        # remove this when no longer needed
        warn_the_user('Extracting may be unsafe; consider updating Python')

Stateful extraction filter example

While tarfile's extraction methods take a simple filter callable, custom filters may be more complex objects with an internal state. It may be useful to write these as context managers, to be used like this:

with StatefulFilter() as filter_func:
    tar.extractall(path, filter=filter_func)

Such a filter can be written as, for example:

class StatefulFilter:
    def __init__(self):
        self.file_count = 0

    def __enter__(self):
        return self

    def __call__(self, member, path):
        self.file_count += 1
        return member

    def __exit__(self, *exc_info):
        print(f'{self.file_count} files extracted')

Command-Line Interface

在 3.4 版新加入.

The tarfile module provides a simple command-line interface to interact with tar archives.

If you want to create a new tar archive, specify its name after the -c option and then list the filename(s) that should be included:

$ python -m tarfile -c monty.tar  spam.txt eggs.txt

Passing a directory is also acceptable:

$ python -m tarfile -c monty.tar life-of-brian_1979/

If you want to extract a tar archive into the current directory, use the -e option:

$ python -m tarfile -e monty.tar

You can also extract a tar archive into a different directory by passing the directory's name:

$ python -m tarfile -e monty.tar  other-dir/

For a list of the files in a tar archive, use the -l option:

$ python -m tarfile -l monty.tar

Command-line options

-l <tarfile>
--list <tarfile>

List files in a tarfile.

-c <tarfile> <source1> ... <sourceN>
--create <tarfile> <source1> ... <sourceN>

Create tarfile from source files.

-e <tarfile> [<output_dir>]
--extract <tarfile> [<output_dir>]

Extract tarfile into the current directory if output_dir is not specified.

-t <tarfile>
--test <tarfile>

Test whether the tarfile is valid or not.

-v, --verbose

Verbose output.

--filter <filtername>

Specifies the filter for --extract. See Extraction filters for details. Only string names are accepted (that is, fully_trusted, tar, and data).

在 3.11.4 版新加入.


How to extract an entire tar archive to the current working directory:

import tarfile
tar = tarfile.open("sample.tar.gz")

How to extract a subset of a tar archive with TarFile.extractall() using a generator function instead of a list:

import os
import tarfile

def py_files(members):
    for tarinfo in members:
        if os.path.splitext(tarinfo.name)[1] == ".py":
            yield tarinfo

tar = tarfile.open("sample.tar.gz")

How to create an uncompressed tar archive from a list of filenames:

import tarfile
tar = tarfile.open("sample.tar", "w")
for name in ["foo", "bar", "quux"]:

The same example using the with statement:

import tarfile
with tarfile.open("sample.tar", "w") as tar:
    for name in ["foo", "bar", "quux"]:

How to read a gzip compressed tar archive and display some member information:

import tarfile
tar = tarfile.open("sample.tar.gz", "r:gz")
for tarinfo in tar:
    print(tarinfo.name, "is", tarinfo.size, "bytes in size and is ", end="")
    if tarinfo.isreg():
        print("a regular file.")
    elif tarinfo.isdir():
        print("a directory.")
        print("something else.")

How to create an archive and reset the user information using the filter parameter in TarFile.add():

import tarfile
def reset(tarinfo):
    tarinfo.uid = tarinfo.gid = 0
    tarinfo.uname = tarinfo.gname = "root"
    return tarinfo
tar = tarfile.open("sample.tar.gz", "w:gz")
tar.add("foo", filter=reset)

Supported tar formats

There are three tar formats that can be created with the tarfile module:

  • The POSIX.1-1988 ustar format (USTAR_FORMAT). It supports filenames up to a length of at best 256 characters and linknames up to 100 characters. The maximum file size is 8 GiB. This is an old and limited but widely supported format.

  • The GNU tar format (GNU_FORMAT). It supports long filenames and linknames, files bigger than 8 GiB and sparse files. It is the de facto standard on GNU/Linux systems. tarfile fully supports the GNU tar extensions for long names, sparse file support is read-only.

  • The POSIX.1-2001 pax format (PAX_FORMAT). It is the most flexible format with virtually no limits. It supports long filenames and linknames, large files and stores pathnames in a portable way. Modern tar implementations, including GNU tar, bsdtar/libarchive and star, fully support extended pax features; some old or unmaintained libraries may not, but should treat pax archives as if they were in the universally supported ustar format. It is the current default format for new archives.

    It extends the existing ustar format with extra headers for information that cannot be stored otherwise. There are two flavours of pax headers: Extended headers only affect the subsequent file header, global headers are valid for the complete archive and affect all following files. All the data in a pax header is encoded in UTF-8 for portability reasons.

There are some more variants of the tar format which can be read, but not created:

  • The ancient V7 format. This is the first tar format from Unix Seventh Edition, storing only regular files and directories. Names must not be longer than 100 characters, there is no user/group name information. Some archives have miscalculated header checksums in case of fields with non-ASCII characters.

  • The SunOS tar extended format. This format is a variant of the POSIX.1-2001 pax format, but is not compatible.

Unicode issues

The tar format was originally conceived to make backups on tape drives with the main focus on preserving file system information. Nowadays tar archives are commonly used for file distribution and exchanging archives over networks. One problem of the original format (which is the basis of all other formats) is that there is no concept of supporting different character encodings. For example, an ordinary tar archive created on a UTF-8 system cannot be read correctly on a Latin-1 system if it contains non-ASCII characters. Textual metadata (like filenames, linknames, user/group names) will appear damaged. Unfortunately, there is no way to autodetect the encoding of an archive. The pax format was designed to solve this problem. It stores non-ASCII metadata using the universal character encoding UTF-8.

The details of character conversion in tarfile are controlled by the encoding and errors keyword arguments of the TarFile class.

encoding defines the character encoding to use for the metadata in the archive. The default value is sys.getfilesystemencoding() or 'ascii' as a fallback. Depending on whether the archive is read or written, the metadata must be either decoded or encoded. If encoding is not set appropriately, this conversion may fail.

The errors argument defines how characters are treated that cannot be converted. Possible values are listed in section Error Handlers. The default scheme is 'surrogateescape' which Python also uses for its file system calls, see File Names, Command Line Arguments, and Environment Variables.

For PAX_FORMAT archives (the default), encoding is generally not needed because all the metadata is stored using UTF-8. encoding is only used in the rare cases when binary pax headers are decoded or when strings with surrogate characters are stored.