Office File Opens as a Folder of XML Files

You double-click a report you saved yesterday, and instead of the document you get a folder containing odd files called [Content_Types].xml, _rels, and word, with document.xml inside it. Sometimes there is no folder — the file opens as a list of XML parts instead.

Nothing was deleted. Word, Excel and PowerPoint files are ZIP archives with a set of XML parts inside, and what you are looking at is the inside of the package. The question is only why Windows stopped handing the file to Office.

Why a .docx looks like a folder

Since Office 2007 every document is an Open Packaging Convention file: a ZIP container whose parts are named, not stored in a fixed order. A .docx holds [Content_Types].xml at the root, a _rels folder of relationship files, word/document.xml with the actual text, word/styles.xml, and word/media/ with the embedded images. An .xlsx uses xl/workbook.xml and xl/worksheets/sheet1.xml; a .pptx uses ppt/presentation.xml and one file per slide under ppt/slides/.

Three things put that packaging on screen:

• The .docx association points at Explorer or an archiver. Windows 10 registered Office extensions to its Compressed Folder handler, so with Word unregistered — a failed install, a portable build, or an Office removal that left the registry behind — double-clicking opens the archive instead of the document. 7-Zip and WinRAR claim the extension the same way.
• You are looking at an extracted copy. Someone renamed report.docx to report.zip, unpacked it, and the resulting folder now sits beside the original. The folder is a copy; the document itself is untouched.
• The name ends in .zip twice. report.xlsx.zip arrives from a mail gateway, a download manager or a colleague on a Mac. File Explorer hides known extensions by default, so the icon reads “report.xlsx” and nothing looks wrong.

Fix 1 — Point the extension back at Word

Right-click the file → Show more options → Open with → Choose another app → Microsoft Word → tick Always use this app. If Word is missing from the list, choose More apps → Look for another app on this PC and browse to C:\Program Files\Microsoft Office\root\Office16\WINWORD.EXE.

For a permanent fix by type, open Settings → Apps → Default apps, type .docx into the search box, and pick Word. Repeat for .xlsx and .pptx — a broken install usually breaks all three at once.

On macOS, select the file, press Cmd+I, set Open with: Microsoft Word, then click Change All.

Fix 2 — Check for a doubled extension

In File Explorer open View → Show → File name extensions. A name reading budget.xlsx.zip is a ZIP archive, and renaming it to budget.xlsx usually works — but only if the archive’s root contains [Content_Types].xml directly. Open it first and look. If you see a single folder name instead, the ZIP wraps a directory and renaming gives you “unreadable content” on the next open.

Fix 3 — Rebuild the package from an extracted folder

When the extracted folder is all you have, zip it back correctly. The parts must sit at the root of the archive; an extra top-level directory makes Word reject the file.

1. Open the folder until [Content_Types].xml is visible in the file list.
2. Ctrl+A to select every item, then right-click → Send to → Compressed (zipped) folder.
3. Rename the new archive from .zip to .docx, .xlsx or .pptx to match the original type.

Compression level and file order do not matter — Office reads parts by name, so a rebuilt archive opens as long as no part is missing. Keep the folder until the rebuilt file opens cleanly, and never re-zip over the original name while testing.

Fix 4 — Pull the text out when the rebuilt file still fails

If the archived file opens with “We found a problem with some content”, the packaging is fine and one of the XML parts is damaged. The text is still readable as raw XML:

1. Open word/document.xml in Notepad or VS Code. Each paragraph of text sits inside <w:t>…</w:t> elements.
2. Copy the whole file into a plain-text editor and strip the tags with a find-and-replace on <[^>]+>. You get the words back, without formatting, tables or images.
3. Rescue images separately from word/media/ — they are ordinary PNG and JPEG files with meaningless names.

For a workbook, the equivalent is xl/worksheets/sheet1.xml for values and xl/sharedStrings.xml for the text they reference.

Fix 5 — Repair the Office registration

If Open with will not keep Word for .docx, the problem is the installation, not the file. Run Settings → Apps → Installed apps → Microsoft 365 → Modify → Quick Repair. It rewrites file associations and registry entries without touching your documents. If associations break again the next day, Online Repair is the next step, at the cost of re-downloading the installation.

When it is not worth rebuilding

The test is whether the text part survived. Open word/document.xml and check its size: a normal page runs to tens of kilobytes. A 0 KB file, or one that ends mid-element, means the document body was never written — rebuild it and Word will complain, and no repair tool can invent the missing words.

The same applies if [Content_Types].xml is absent (nothing can open the package), or if the folder contains only _rels and docProps with no word directory. In an .xlsx, a sheet1.xml of a few hundred bytes with no <sheetData> is an empty sheet, not a damaged one.

What is usually salvageable even then: images from word/media/, and any text you can still read in document.xml. If the body is truncated halfway, expect roughly the first half of the document as plain text and nothing after it. If you would rather not dig through XML parts yourself, send the file over — the parts tell us quickly whether the text is there.

FAQ

Is my file corrupted if I can see the XML?

No. Seeing the parts is normal — every Office file is built from them. It means the shell opened the container instead of passing it to Word, or that you are looking at an extracted copy. Corrupted means a part is missing or malformed, which is a different check.

Can I just rename the folder to .docx?

No. A folder is not an archive; the rename either fails or produces a file Office cannot read. Compress the contents first, as in Fix 3.

Does the same thing happen with .pptx and .xlsx?

Yes, and the internals follow the same pattern. Only the part names change. A .pptx whose ppt/slides/ folder is empty has lost its slides, while its text placeholders may still exist in the slide layouts.