This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

Title: Improve the ZipFile Interface
Type: enhancement Stage:
Components: Library (Lib) Versions:
Status: closed Resolution: accepted
Dependencies: Superseder:
Assigned To: Nosy List: alanmcintyre, crhode, georg.brandl, jvr, myers_carpenter, mzimmerman, scott_daniels
Priority: normal Keywords:

Created on 2001-10-04 15:54 by anonymous, last changed 2022-04-10 16:04 by admin. This issue is now closed.

File name Uploaded Description Edit
zipfile_extract.diff alanmcintyre, 2008-01-05 00:30 Add extract/extractall methods to ZipFile
Messages (12)
msg53289 - (view) Author: Nobody/Anonymous (nobody) Date: 2001-10-04 15:54
There exist two methods to write to a ZipFile

     write(self, filename, arcname=None, compress_type=None)  
     writestr(self, zinfo, bytes)

but only one to read from it

     read(self, name)

Additionally, the two 'write's behave differently with respect to compression.

(a) 'read' does not fit to 'write', since 'write' takes a file and adds it to a ZipFile, 
     but 'read' is not the reverse operation. 'read' should be called 'readstr' since it 
     much better matches to 'writestr'.

(b) It is confusing what 'write' and 'read' actually mean. Does 'write' write a file, 
     or into the ZipFile? It would be more obvious if ZipFile has 4 methods which 
     pair-wise fit together:

     writestr (self, zinfo, bytes)
          # same as now
     readstr (self, name)
          # returns bytes (as string), currently called 'read'
          # 'read' could still live but should be deprecated
     add (self, filename, arcname=None, compress_type=None)
          # currently 'write'
          # 'write' could still live but should be deprecated
     extract (self, name, filename, arcname=None)
          # new, desired functionality

(c) BOTH, 'writestr' and 'add' should by default use the 'compress_type' that was 
     passed to the constructor of 'ZipFile'. Currently, 'write' does it, 'writestr' via 
     zinfo does it not. 'ZipInfo' sets the compression strict to 'ZIP_STORED' :-( 
     It should not do that! It rather should:
     - allow more parameters in the signature of the constructor
        to also pass the compression type (and some other attributes, too)
     - default to 'None', so that 'writestr' can see this, and then take 
        the default from the 'ZipFile' instance.

msg53290 - (view) Author: Just van Rossum (jvr) * (Python triager) Date: 2003-01-05 20:54
Logged In: YES 

In Python 2.3, writestr() has an enhanced signature: the
first arg may now also be an archive name, in which case the
correct default settings are used (ie. the compression value
is taken from the file). See patch #651621.

extract() could be moderately useful (although I don't
understand the 'arcname' arg, how's that different from
'name'?) but would have to deal with file modes (bin/text).
The file mode isn't in the archive so would have to
(optionally) be supplied by the caller.
msg53291 - (view) Author: Matt Zimmerman (mzimmerman) Date: 2003-07-31 14:22
Logged In: YES 

It would also be very useful to be able to have ZipFile
read/write the uncompressed file data from/to a file-like
object, instead of just strings and files (respectively).

I would like to use this module to work with zip files
containing large files, but this is unworkable because the
current implementation would use excessive amounts of memory.

Currently, read() reads all of the compressed data into
memory, then uncompresses it into memory.  For files which
may be hundreds of megabytes compressed, this is undesirable.

Likewise for write(), I would like to be able to stream data
into a zip file, passing in a ZipInfo to specify the
metadata as is done with writestr().

The implementation of this functionality is quite
straightforward, but I am not sure whether (or how) the
interface should change.  Some other parts of the library
allow for a file object to be passed to the same interface
which accepts a filename.  The object is examined to see if
it has the necessary read/write methods and if not, it is
assumed to be a filename.  Would this be the correct way to
do it?

I, too, am a bit irked by the lack of symmetry exhibited by
read vs. write/writestr, and think that the interface
suggested above would be a significant improvement.
msg53292 - (view) Author: Myers Carpenter (myers_carpenter) Date: 2004-05-09 18:23
Logged In: YES 

The zipfile interface should match the tarfile interface. 
At the mininum is should work for this example:

import zipfile
zip ="", "r")
for zipinfo in zip:
    print, "is", tarinfo.size, "bytes in size
and is",

This closely matchs the 'tarfile' module.
msg53293 - (view) Author: Chuck Rhode (crhode) Date: 2005-09-22 15:56
Logged In: YES 

I've been trying to read map files put out by the Census
Bureau.  These ZIP archives are downloaded from government
contractors' sites by county.  Within each county archive
are several ZIP files for each map layer (roads, streams,
waterbodies, etc).  Each contains the elements of an ESRI
shapefile database (.shp, .shx., and .dbf files).  This
doesn't make a lot of sense to me, either, because there's
no compression advantage to making an archive of an archive.
 The technique is used purely for organizational purposes
because ZIP does not compress subdirectories.

Note: I've never seen a TAR of TAR files because TAR *does*
compress subdirectories.

What I've been struggling with is a way to leave these
archives in their compressed form and still do *python* I/O
on them.  There is a tree organization to them, after all,
just as with traditional os.path directories.  I've designed
some objects that let me retrieve the most recent file, ZIP
member, or TAR member by name from a given path to a
repository of such archives.  What I get is a StreamIO
object that I can subsequently put back where it came from.

What would be nice is if there already were objects
available to manipulate normal os.path directories comingled
with ZIP and TAR archives.  What would be nicer is if I/O
could be opened at the character/line level transparently
without regard to whether the source/destination was a file
or an archive member within such a structure.  In the days
of hardware compression and on-the-fly encryption/decryption
of I/O, is this too much to ask?  -ccr-
msg53294 - (view) Author: Scott David Daniels (scott_daniels) * Date: 2005-09-25 20:20
Logged In: YES 

I am currently working on an expanded zipfile module that:
(a) Has a more easily extensible class
(b) Allows BZIP2 compression (my orginal need)
(c) Allows file-like (read) access to the elements of ZipFile
(d) Provides for a single "writer" which can be used to 
     generate file contents "incrementally" while possibly
     reading from other "files" in the zipfile
(e) Allows the opening of embedded zips "in-place"

What I don't have at the moment is a good set of tests
or good documents of how to use it.  Anyone interested 
in collaborating, let me know.

--Scott David Daniels
msg53295 - (view) Author: Alan McIntyre (alanmcintyre) * (Python committer) Date: 2005-09-25 23:29
Logged In: YES 


I had put together some enhancements to ZipFile read/write,
including test cases, but haven't had time to advocate
getting it into 2.5.  You can find it here:

If it seems like it would be helpful, I can go round up the
most recent version (that I've been using in a production
environment) and send it to you.
msg57981 - (view) Author: Georg Brandl (georg.brandl) * (Python committer) Date: 2007-11-30 13:56
Alan's patch has since been committed. Is there any more work on this item?
msg57990 - (view) Author: Alan McIntyre (alanmcintyre) * (Python committer) Date: 2007-11-30 16:21
There was another issue that also asked for an extract feature, and if I
recall correctly I said I'd try to work on it (I think I have some code
somewhere for it but I'll have to look).  Tonight or tomorrow I will see
if I can find that other issue and let you know about it, and maybe take
a look around at the various zipfile improvement/change requests to see
if they've been completely addressed.
msg59269 - (view) Author: Alan McIntyre (alanmcintyre) * (Python committer) Date: 2008-01-05 00:30
I attached a patch with the following changes (as zipfile_extract.diff):

(1) Add a note to the docs (under writestr) about how the compression is
selected if a ZipInfo is passed as the zinfo_or_arcname parameter.  If
anybody thinks it's a good idea to add a compression argument to the
ZipInfo constructor, I can modify the patch/docs accordingly.

(2) Add an extract method to ZipFile and associated test/documentation
msg59272 - (view) Author: Alan McIntyre (alanmcintyre) * (Python committer) Date: 2008-01-05 00:39
Are the method renames/additions suggested in the original issue worth
doing?  When I first started using this module, I found the
documentation easy and thorough enough to understand how to use it, so I
would vote for just leaving the ZipFile interface the way it is.
msg59476 - (view) Author: Georg Brandl (georg.brandl) * (Python committer) Date: 2008-01-07 18:48
I committed your patch (after reviewing the docs) as r59834. I think
there is no more to do here.
Date User Action Args
2022-04-10 16:04:30adminsetgithub: 35279
2008-01-07 18:48:55georg.brandlsetstatus: open -> closed
resolution: accepted
messages: + msg59476
2008-01-05 00:39:26alanmcintyresetmessages: + msg59272
2008-01-05 00:30:54alanmcintyresetfiles: + zipfile_extract.diff
messages: + msg59269
2007-11-30 16:21:47alanmcintyresetmessages: + msg57990
2007-11-30 13:56:06georg.brandlsetnosy: + georg.brandl
messages: + msg57981
2001-10-04 15:54:14anonymouscreate