| {"cwd": "/testbed", "dataset_type": "opensource-code", "docker_image": "format-code-task-002220:latest", "instance_id": "format-code-task-002220", "problem_statement": "BUG: ExcelWriter with mode='a' corrupts file\n- [x] I have checked that this issue has not already been reported.\r\n\r\n- [x] I have confirmed this bug exists on the latest version of pandas.\r\n\r\n- [ ] (optional) I have confirmed this bug exists on the master branch of pandas.\r\n\r\n---\r\n\r\n**Note**: Please read [this guide](https://matthewrocklin.com/blog/work/2018/02/28/minimal-bug-reports) detailing how to provide the necessary information for us to reproduce your bug.\r\n\r\n#### Code Sample, a copy-pastable example\r\n\r\n```python\r\nimport pandas as pd\r\n\r\ndf = pd.DataFrame()\r\n\r\nwith pd.ExcelWriter('test.xlsx', mode='w', engine='openpyxl') as writer:\r\n df.to_excel(writer)\r\n\r\nwith pd.ExcelWriter('test.xlsx', mode='a', engine='openpyxl') as writer:\r\n df.to_excel(writer)\r\n\r\n```\r\n\r\n#### Problem description\r\n\r\nAppending to excel file with pandas>=1.2 produces file which Excel reports as corrupted and opens only after repair, forcing you to first open and repair file before passing it to someone else.\r\nLooking at file with 7-Zip File Manager shows that after writing file with mode='a', some .xml files (app.xml, core.xml, theme1.xml, sheet1.xml) inside archive are duplicated (somehow). After Excel rapars file duplicates are gone, everything else looks the same.\r\n\r\nWith pandas 1.1.5 and same versions of other packages everything works as expected.\r\n\r\n#### Expected Output\r\n\r\nFile which could be opened by Excel without any warnings.\r\n\r\n#### Output of ``pd.show_versions()``\r\n\r\n<details>\r\n\r\nINSTALLED VERSIONS\r\n------------------\r\ncommit : 9d598a5e1eee26df95b3910e3f2934890d062caa\r\npython : 3.8.6.final.0\r\npython-bits : 64\r\nOS : Windows\r\nOS-release : 10\r\nVersion : 10.0.19041\r\nmachine : AMD64\r\nprocessor : Intel64 Family 6 Model 158 Stepping 10, GenuineIntel\r\nbyteorder : little\r\nLC_ALL : None\r\nLANG : None\r\nLOCALE : Russian_Russia.1251\r\npandas : 1.2.1\r\nnumpy : 1.20.0\r\npytz : 2021.1\r\ndateutil : 2.8.1\r\npip : 21.0.1\r\nsetuptools : 49.6.0.post20210108\r\nCython : None\r\npytest : None\r\nhypothesis : None\r\nsphinx : None\r\nblosc : None\r\nfeather : None\r\nxlsxwriter : None\r\nlxml.etree : 4.6.2\r\nhtml5lib : None\r\npymysql : None\r\npsycopg2 : None\r\njinja2 : None\r\nIPython : None\r\npandas_datareader: None\r\nbs4 : None\r\nbottleneck : None\r\nfsspec : None\r\nfastparquet : None\r\ngcsfs : None\r\nmatplotlib : None\r\nnumexpr : None\r\nodfpy : None\r\nopenpyxl : 3.0.6\r\npandas_gbq : None\r\npyarrow : None\r\npyxlsb : None\r\ns3fs : None\r\nscipy : None\r\nsqlalchemy : None\r\ntables : None\r\ntabulate : None\r\nxarray : None\r\nxlrd : None\r\nxlwt : None\r\nnumba : None\r\n\r\n</details>", "test_command": "bash /testbed/mimo_test_command.sh", "test_patch": "diff --git a/pandas/tests/io/excel/test_openpyxl.py b/pandas/tests/io/excel/test_openpyxl.py\nindex 640501baff..cc22795a76 100644\n--- a/pandas/tests/io/excel/test_openpyxl.py\n+++ b/pandas/tests/io/excel/test_openpyxl.py\n@@ -149,3 +149,100 @@ def test_read_with_bad_dimension(datapath, ext, header, expected_data, filename)\n result = pd.read_excel(path, header=header)\n expected = DataFrame(expected_data)\n tm.assert_frame_equal(result, expected)\n+\n+\n+def test_append_mode_file_not_corrupted_empty_df(ext):\n+\n+ import os\n+\n+ df = DataFrame()\n+ with tm.ensure_clean(ext) as f:\n+ with ExcelWriter(f, mode=\"w\", engine=\"openpyxl\") as writer:\n+ df.to_excel(writer)\n+ size_after_w = os.path.getsize(f)\n+\n+ with ExcelWriter(f, mode=\"a\", engine=\"openpyxl\") as writer:\n+ df.to_excel(writer, sheet_name=\"Sheet2\")\n+ size_after_a = os.path.getsize(f)\n+\n+\n+\n+\n+ assert size_after_a < size_after_w * 1.5, (\n+ f\"File bloated after append: {size_after_w} -> {size_after_a}\"\n+ )\n+\n+\n+def test_append_mode_file_size_matches_fresh_write(ext):\n+\n+\n+ import os\n+\n+ df = DataFrame({\"A\": [1, 2, 3]})\n+\n+ with tm.ensure_clean(ext) as f_appended:\n+ with ExcelWriter(f_appended, mode=\"w\", engine=\"openpyxl\") as writer:\n+ df.to_excel(writer, sheet_name=\"S1\")\n+ with ExcelWriter(f_appended, mode=\"a\", engine=\"openpyxl\") as writer:\n+ df.to_excel(writer, sheet_name=\"S2\")\n+ size_appended = os.path.getsize(f_appended)\n+\n+ with tm.ensure_clean(ext) as f_fresh:\n+ with ExcelWriter(f_fresh, mode=\"w\", engine=\"openpyxl\") as writer:\n+ df.to_excel(writer, sheet_name=\"S1\")\n+ df.to_excel(writer, sheet_name=\"S2\")\n+ size_fresh = os.path.getsize(f_fresh)\n+\n+\n+ assert size_appended < size_fresh * 1.2, (\n+ f\"Append-mode file larger than expected: fresh={size_fresh}, \"\n+ f\"appended={size_appended}\"\n+ )\n+\n+\n+def test_append_mode_repeated_appends_no_bloat(ext):\n+\n+ import os\n+\n+ df = DataFrame({\"A\": [1]})\n+ with tm.ensure_clean(ext) as f:\n+ with ExcelWriter(f, mode=\"w\", engine=\"openpyxl\") as writer:\n+ df.to_excel(writer, sheet_name=\"S0\")\n+ size_initial = os.path.getsize(f)\n+\n+ for i in range(1, 4):\n+ with ExcelWriter(f, mode=\"a\", engine=\"openpyxl\") as writer:\n+ df.to_excel(writer, sheet_name=f\"S{i}\")\n+\n+ size_final = os.path.getsize(f)\n+\n+\n+ with tm.ensure_clean(ext) as f_fresh:\n+ with ExcelWriter(f_fresh, mode=\"w\", engine=\"openpyxl\") as writer:\n+ for i in range(4):\n+ df.to_excel(writer, sheet_name=f\"S{i}\")\n+ size_fresh = os.path.getsize(f_fresh)\n+\n+\n+\n+ assert size_final < size_fresh * 1.5, (\n+ f\"Repeated appends bloated file: fresh={size_fresh}, \"\n+ f\"after_appends={size_final}\"\n+ )\n+\n+\n+def test_append_mode_data_readable(ext):\n+\n+ df1 = DataFrame({\"A\": [1, 2, 3]})\n+ df2 = DataFrame({\"B\": [4, 5, 6]})\n+\n+ with tm.ensure_clean(ext) as f:\n+ with ExcelWriter(f, mode=\"w\", engine=\"openpyxl\") as writer:\n+ df1.to_excel(writer, sheet_name=\"S1\", index=False)\n+ with ExcelWriter(f, mode=\"a\", engine=\"openpyxl\") as writer:\n+ df2.to_excel(writer, sheet_name=\"S2\", index=False)\n+\n+ sheets = pd.read_excel(f, sheet_name=None)\n+ assert set(sheets.keys()) == {\"S1\", \"S2\"}\n+ tm.assert_frame_equal(sheets[\"S1\"], df1)\n+ tm.assert_frame_equal(sheets[\"S2\"], df2)\ndiff --git a/test_commands.json b/test_commands.json\nnew file mode 100644\nindex 0000000000..f802b9870d\n--- /dev/null\n+++ b/test_commands.json\n@@ -0,0 +1,3 @@\n+{\n+ \"test_commands\": [\"python -m pytest pandas/tests/io/excel/test_openpyxl.py -k 'append_mode_file_not_corrupted_empty_df or append_mode_file_size_matches_fresh_write or append_mode_repeated_appends_no_bloat or append_mode_data_readable' -v\"]\n+}\ndiff --git a/mimo_test_command.sh b/mimo_test_command.sh\nnew file mode 100755\n--- /dev/null\n+++ b/mimo_test_command.sh\n@@ -0,0 +1,8 @@\n+#!/usr/bin/env bash\n+\n+set -uo pipefail\n+cd /testbed\n+rc=0\n+python -m pytest pandas/tests/io/excel/test_openpyxl.py -k 'append_mode_file_not_corrupted_empty_df or append_mode_file_size_matches_fresh_write or append_mode_repeated_appends_no_bloat or append_mode_data_readable' -v\n+rc=$(( rc | $? ))\n+exit $rc\n", "verifier_timeout_sec": 1800} |