From fb353446651067fafec03301256e93d59065c520 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Torsten=20W=C3=B6rtwein?=
 <twoertwein@users.noreply.github.com>
Date: Mon, 28 Dec 2020 19:05:06 -0500
Subject: [PATCH] REGR: to_csv created corrupt ZIP files when there are more
 rows than chunksize (#38728)

---
 doc/source/whatsnew/v1.2.1.rst         |  1 +
 pandas/io/common.py                    | 29 +++++++++++++++++++++-----
 pandas/tests/io/formats/test_to_csv.py | 22 +++++++++++++++++++
 3 files changed, 47 insertions(+), 5 deletions(-)

diff --git a/doc/source/whatsnew/v1.2.1.rst b/doc/source/whatsnew/v1.2.1.rst
index 769c195229bbd..e280b730679f0 100644
--- a/doc/source/whatsnew/v1.2.1.rst
+++ b/doc/source/whatsnew/v1.2.1.rst
@@ -15,6 +15,7 @@ including other versions of pandas.
 Fixed regressions
 ~~~~~~~~~~~~~~~~~
 - The deprecated attributes ``_AXIS_NAMES`` and ``_AXIS_NUMBERS`` of :class:`DataFrame` and :class:`Series` will no longer show up in ``dir`` or ``inspect.getmembers`` calls (:issue:`38740`)
+- :meth:`to_csv` created corrupted zip files when there were more rows than ``chunksize`` (issue:`38714`)
 -
 
 .. ---------------------------------------------------------------------------
diff --git a/pandas/io/common.py b/pandas/io/common.py
index c33ef9ac4ba95..642684ca61480 100644
--- a/pandas/io/common.py
+++ b/pandas/io/common.py
@@ -4,10 +4,10 @@
 from collections import abc
 import dataclasses
 import gzip
-from io import BufferedIOBase, BytesIO, RawIOBase, TextIOWrapper
+from io import BufferedIOBase, BytesIO, RawIOBase, StringIO, TextIOWrapper
 import mmap
 import os
-from typing import IO, Any, AnyStr, Dict, List, Mapping, Optional, Tuple, cast
+from typing import IO, Any, AnyStr, Dict, List, Mapping, Optional, Tuple, Union, cast
 from urllib.parse import (
     urljoin,
     urlparse as parse_url,
@@ -713,17 +713,36 @@ def __init__(
         archive_name: Optional[str] = None,
         **kwargs,
     ):
-        if mode in ["wb", "rb"]:
-            mode = mode.replace("b", "")
+        mode = mode.replace("b", "")
         self.archive_name = archive_name
+        self.multiple_write_buffer: Optional[Union[StringIO, BytesIO]] = None
+
         kwargs_zip: Dict[str, Any] = {"compression": zipfile.ZIP_DEFLATED}
         kwargs_zip.update(kwargs)
+
         super().__init__(file, mode, **kwargs_zip)  # type: ignore[arg-type]
 
     def write(self, data):
+        # buffer multiple write calls, write on flush
+        if self.multiple_write_buffer is None:
+            self.multiple_write_buffer = (
+                BytesIO() if isinstance(data, bytes) else StringIO()
+            )
+        self.multiple_write_buffer.write(data)
+
+    def flush(self) -> None:
+        # write to actual handle and close write buffer
+        if self.multiple_write_buffer is None or self.multiple_write_buffer.closed:
+            return
+
         # ZipFile needs a non-empty string
         archive_name = self.archive_name or self.filename or "zip"
-        super().writestr(archive_name, data)
+        with self.multiple_write_buffer:
+            super().writestr(archive_name, self.multiple_write_buffer.getvalue())
+
+    def close(self):
+        self.flush()
+        super().close()
 
     @property
     def closed(self):
diff --git a/pandas/tests/io/formats/test_to_csv.py b/pandas/tests/io/formats/test_to_csv.py
index a9673ded7c377..6416cb93c7ff5 100644
--- a/pandas/tests/io/formats/test_to_csv.py
+++ b/pandas/tests/io/formats/test_to_csv.py
@@ -640,3 +640,25 @@ def test_to_csv_encoding_binary_handle(self, mode):
 
                 handle.seek(0)
                 assert handle.read().startswith(b'\xef\xbb\xbf""')
+
+
+def test_to_csv_iterative_compression_name(compression):
+    # GH 38714
+    df = tm.makeDataFrame()
+    with tm.ensure_clean() as path:
+        df.to_csv(path, compression=compression, chunksize=1)
+        tm.assert_frame_equal(
+            pd.read_csv(path, compression=compression, index_col=0), df
+        )
+
+
+def test_to_csv_iterative_compression_buffer(compression):
+    # GH 38714
+    df = tm.makeDataFrame()
+    with io.BytesIO() as buffer:
+        df.to_csv(buffer, compression=compression, chunksize=1)
+        buffer.seek(0)
+        tm.assert_frame_equal(
+            pd.read_csv(buffer, compression=compression, index_col=0), df
+        )
+        assert not buffer.closed