Python ZipFile Created In-Memory Not Compressing as Expected

Question

I am trying to use Python to create a ZipFile object in-memory, and write a single file, also created in-memory, into the ZipFile object, and then upload the file to Google Cloud Storage.

My file is not actually getting compressed. Any idea what I might be doing wrong?

I realize there may be a fancier way of getting the row data into the file object, but apart from that, I'm really just trying to figure out why the resulting zip file is not coming out compressed at all.

UPDATE: code sample now excludes any interaction with Google Cloud Services (GCS, etc.), and instead just writes the files to disk.

It seems that when I write the file to disk first, then create the ZipFile, the result is compressed as expected, but when I add the StringIO contents directly from memory to the ZipFile object, the contents are not compressed.

import random, io, argparse, os, string
from zipfile import ZipFile, ZipInfo, ZIP_DEFLATED

parser = argparse.ArgumentParser()
parser.add_argument("--row_limit", default=1000)
parser.add_argument("--file_name", default='file.txt', type=str)
parser.add_argument("--archive_name", default='file.zip', type=str)
parser.add_argument("--snapshot_millis", default=0, type=int)
args = parser.parse_args()

# imagine this has lots and lots of data in it, coming from a database query result
rows = [{
    'seq_no': ''.join(random.choices(string.ascii_uppercase + string.digits, k=args.row_limit)),
    'csv': ''.join(random.choices(string.ascii_uppercase + string.digits, k=args.row_limit))
}] * args.row_limit

archive = io.BytesIO()
# create zip archive in memory
with ZipFile(archive, 'w', compression=ZIP_DEFLATED, compresslevel=9) as zip_archive:
    count = 0
    file_contents = io.StringIO()
    for row in rows:
        if count > args.row_limit:
            break
        count += 1
        file_contents.write(f"{row['seq_no']},{row['csv']}\n")

    # write file to zip archive in memory
    zip_file = ZipInfo(args.file_name)
    zip_archive.writestr(zip_file, file_contents.getvalue())

    # also write file to disk
    with open(args.file_name, mode='w') as f:
        print(file_contents.getvalue(), file=f)

    print(f"StringIO Size: {file_contents.tell()}")
    print(f"Text File Size On Disk: {os.path.getsize(args.file_name)}")

archive.seek(0)

with open(args.archive_name, 'wb') as outfile:
    outfile.write(archive.getbuffer())

print(f"Zip File Created from File In Memory: {os.path.getsize(args.archive_name)}")

ZipFile(args.archive_name, mode='w', compression=ZIP_DEFLATED, compresslevel=9).write(args.file_name)

print(f"Zip File Created from File On Disk: {os.path.getsize(args.archive_name)}")

Does the upload to the cloud storage makes any difference? If not, please edit your code to demonstrate the problem with the zipfile (and nothing else). — VPfB, Apr 06 '22 at 05:32
Thanks, I updated the code sample to run completely on its own, without any connection to Google Cloud Services. GCS doesn't seem to have an impact on the compression issue. — Dave Morris, Apr 06 '22 at 19:30

VPfB · Accepted Answer · 2022-04-07T06:14:32.703

The problem is here:

zip_file = ZipInfo(args.file_name)
zip_archive.writestr(zip_file, file_contents.getvalue())

From the ZipFile.writestr docs:

When passing a ZipInfo instance as the zinfo_or_arcname parameter, the compression method used will be that specified in the compress_type member of the given ZipInfo instance. By default, the ZipInfo constructor sets this member to ZIP_STORED [i.e. uncompressed].

The easiest way to correct the issue is not to use a complete ZipInfo, but just the file name. This will also set the current date/time as the creation time for the file inside the archive (ZipInfo defaults to year 1980):

# zip_file = ZipInfo(args.file_name)
zip_archive.writestr(args.file_name, file_contents.getvalue())

Thank you, that seems to be working perfectly for me! – Dave Morris Apr 06 '22 at 22:26 — Dave Morris, Apr 06 '22 at 22:26

Python ZipFile Created In-Memory Not Compressing as Expected

1 Answers1