How can I recover a corrupt document or template – and why did it become corrupt?

Article contributed by Dave Rado and John McGhie

The advice in this article applies to all versions of Word, including those for the Macintosh.

Why do documents corrupt?

If you use any of the following features, your documents are likely to corrupt: Master Documents, Nested Tables, Versions, Fast Saves, Document Map, and saving to a floppy. For more on these, see: Tips and “Gotchas”.

In addition, saving when resources are low can cause corruptions. If you notice Word start to slow down noticeably it's always best to quit and restart Word immediately; to close any other applications that are open; and to clear the clipboard, by selecting any character and copying it.

Other signs that you are low on resources: fonts suddenly not displaying properly; the wrong application icons appearing on your Desktop or in Windows Explorer (e.g. Word's icon appearing where Excel's should be). If you get these symptoms, restart Windows immediately.

A corrupt printer driver can corrupt memory; and if you then save, this can corrupt the document. Symptoms: Word often crashes when printing (cure: reinstall the driver).

A corrupt template will corrupt any new documents based on it. A corrupt Normal.dot template is especially bad -- it spreads its evil almost like a virus to almost every new document you create.

If you create list numbering using the Format + Bullets and Numbering dialog, this is likely to lead to a corruption arising eventually, especially if you also have the “Automatically update document styles” checkbox ticked on the Tools + Templates and Addins dialog (less of a problem with Heading numbering than other types).

A bad sector on your hard disk can corrupt a document saved to that sector. Make sure you run Scandisk regularly. Running Defrag regularly will also help reduce the chances of running low on resources. Encourage users to save to the network. Make regular backups.

Where are corruptions stored?

Corruptions are usually, but not always, stored in Section Breaks. The final paragraph mark in a document contains a hidden Section Break, so in a single-section document, corruptions tend to be stored in the final paragraph mark.

Corruptions can also be stored in any paragraph mark in a document; or in an end-of-cell or end-of-row marker within a table; or in a bookmark. (Corrupt bookmarks are very rare in Word 97 and above, unless you have been using EndNote).

If you find that certain commands such as Edit>Find don't work within a certain table, that table is probably corrupt.

If you find text mysteriously disappearing and reappearing as you page down past a particular paragraph, that paragraph's paragraph mark is likely to be corrupt (see the section on fixing corrupt templates).

How can I fix a corrupt document?

If you have been using Master documents, see How to recover a Master Document.

If all new documents based on a certain template are showing symptoms of corruption, the template they are based on is almost certainly corrupt.

Otherwise:

If using Word 2000 or Above:

Select File + Save As Web page, quit Word, reopen the htm file and save it back in Word format – that usually (but not always) gets rid of corruptions. The HTML/XML format forces Word to completely re-create the internal structure of the document, either fixing or discarding the corrupt bits when it does.  Best of all, in the case of Word 2000 and above, almost all of the formatting and page layout is preserved.

Please note: to preserve your formatting, you must select the plain Save As "Web Page" option, not the Save As Web Page (Filtered).  If you use the Filtered option, you remove from the document all the formatting that an HTML browser cannot interpret: for example, page numbers and headers and footers!

If that doesn't fix it, the fixes described below apply.

If using Word 97 or above:

  1. If you have isolated the corruption to a particular table, either:
    • Paste the table into Excel; delete the Word table; paste the Excel table back into Word, select the new table (Alt+Double-click), press Ctrl+Spacebar to remove the manual formatting, and reformat the table, or:
    • Select Table + Convert Table to Text, select the text that results, and select Table + Convert Text to Table. This has the advantage that you lose much less formatting than using the Excel method, but the disadvantage that if a corruption is stored in a paragraph mark within the table, it will remain.
  2. If the table contains horizontally merged cells, it's best to recreate a few rows at a time – for instance, if using method (b), then after converting the table to text, select contiguous rows that have equal numbers of columns, convert them to a table, and keep doing this until you have converted all the text to individual tables (which will automatically merge themselves into a single table).
  3. If you have isolated the corruption to a particular paragraph, select the text in the paragraph, but be careful not to select the paragraph marker (the paragraph marker is a property container, and that's where the corruption is stored). Paste into a new document. Delete the corrupt paragraph. Paste back from the new document to the old one.
  4. You can try saving as RTF, closing Word, reopening the RTF file and saving back as a Word document. Unfortunately, Word's RTF format is similar enough to Word's native format to preserve most corruptions.
  5. If that doesn't work, delete any Section Breaks using Find and Replace, then Select All (Ctrl+A), de-select the final paragraph mark (Shift + Left arrow), copy, and paste into a new document. Then close the corrupt document and save the new one, overwriting the old one (in that order). Finally, log out or restart your operating system  before doing anything more (because document corruptions can corrupt memory).
  6. If even that doesn't work, try saving in Word 2 format if you have this option (the Word 2 converter is no longer offered, but if you have upgraded from a previous version, you will still have it). Unfortunately, you will lose an awful lot of formatting if you do this, though.
  7. If even that doesn't work, select File + Open, set the “Files of type” list box to “Recover Text from Any File”, and open the corrupt document. Delete the gobbledygook at the end. You'll lose all of your formatting leaving only the text.

Note that the “Recover Text from Any File” setting is “sticky”. In Word versions prior to 2002, you must immediately select File + Open again, change the  “Files of type” setting back to “Word Document” and open another document, while you remember.  If you forget, every file you open will have no formatting, and if you save it in this condition, it's gone forever.  See Whenever I open a document using File Open all my formatting is gone, and there is garbage at the end).

How can I fix a corrupt template?

The best strategy is to keep a backup, macro-free version of all your templates. Then if a template become corrupt, you can copy any macros over to a copy of the backup template using the Organizer and you're back in business.

If you haven't done that, though, and if your template contains any macros, you could try running the VBA Code Cleaner.

If that doesn't fix it, recreate the template from scratch. If you want to copy the content from the corrupt template to the new one, follow the same steps as for fixing a corrupt document. It may be worth creating the new template based on a “virgin” copy of Normal.dot, just in case Normal.dot is also corrupt. With Word closed, rename your existing Normal.dot file and restart Word; a new copy of Normal.dot will automatically be created.

One more thing; under Tools + Options + Save, turn on the checkbox which says “Prompt to save Normal template”, if it isn't switched on already (unfortunately, it is switched off by default). The only time you should ever save Normal.dot is when you have knowingly made a change to it that you want to save. Then you should save it by holding the Shift key down and selecting File + Save All. And be sure never to save Normal.dot when resources are low (see: Why do documents corrupt?).

Allowing Word to save Normal.dot whether you've consciously made changes to it or not, is OK in versions of Word later than Word 97/98.