# Fix python-docx "KeyError: document.xml" opening a counterparty .docx

## TL;DR
The file isn't a valid Office Open XML package: it's usually a legacy .doc that was renamed to .docx, or a producer wrote a zip missing word/document.xml. Check the zip contents to confirm, then open the file in Word or LibreOffice and save it as a real .docx; python-docx will open the repaired file.

```text
KeyError: 'document.xml'
```

## Steps
1. Inspect the package: list the zip contents and look for word/document.xml.
   Expected: the entry is missing (or the file isn't a zip at all, meaning it's a renamed legacy .doc).
2. Open the file in Word or LibreOffice and save it as .docx (Office Open XML).
   Expected: the saved file is a proper package with word/document.xml present.
3. For headless batch repair: convert with LibreOffice headless to docx.
   Expected: a clean output file that python-docx opens.
4. Re-run your python-docx code against the repaired file.
   Expected: the Document opens with paragraphs and tables accessible.

## Use this when
- python-docx raises KeyError: 'document.xml' on a .docx from a counterparty
- The file opens in Word but not in python-docx
- Files from non-Word producers (exports, converters) fail to open

## Not for this skill when
- python-docx opens the file but content looks wrong (that's a content problem, not a package problem)
- The file is password-protected (decrypt first; Office encryption changes the package layout)
- You're writing .docx files and they come out corrupt (that's a generation problem)

## Variant phrasings
### python-docx keyerror document xml
The package is malformed; re-save through Word or LibreOffice.
### docx file is not a valid zip python
Often a renamed .doc; convert rather than renaming back.
### open corrupt docx python-docx
Repair the package first; python-docx can't parse non-packages.

## Why it happens
A .docx is a zip containing a fixed set of XML parts, and python-docx expects word/document.xml among them. Renamed legacy .doc files are a different binary format entirely, and some exporters write zips with non-standard layouts. Word tolerates both; python-docx does not.

## Edge cases
- Google Docs and modern exporters usually produce valid packages; suspect the file's origin when they don't.
- A renamed .doc needs real conversion, not renaming back; the formats are unrelated.
- Macros (.docm) renamed to .docx can also confuse strict readers; keep the right extension.

## Provenance

Resolved from the public thread: https://vectle.com/posts/pst_ope_-XVw6fJTO5McUHBNDw