Permanently remove Collected Words associated with Writer documents

Windows 10 Pro – LibreOffice 7.2.0.4 64-bit - Writer .docx and .odt documents.

I wish to clear the entire list of autocorrect (user collected) entries. Tools, AutoCorrect, AutoCorrect Options, Word Completion, (Collected Words this list).

If I delete every item in the above list, they simply return the next time a document is opened. This is true if I clear the list with the document that added them, or if I delete the entries while in a blank document.

If I delete every acor_en-US, en-GB etc .dat file from C:\Users\UserName\AppData\Roaming\LibreOffice\4\user\autocorr folder, or from C:\Program Files\LibreOffice\share\autocorr, the Collected Words list gets repopulated the next time I open a document.

If I delete my Profile 4 folder and start with a blank default 4 folder, the list is again repopulated the next time I open a document.

Does each document carry its own embedded word list, and the reason why the list repopulates with a new Profile 4 folder? If yes, how can I permanently remove collected words? Whatever I try, the same items reappear. If I copy and paste text to a blank document, the Collected Words, from the source document, appear in the list of the new document.

Please note, my question is specific to Word Completion, Collected Words, not AutoCorrect in general.

I have searched the community, but found nothing about permanently removing Collected Words. Or at least found nothing that worked.

Thank you in advance for any assistance.

You mention in your question you work with .docx documents.

.docx documents are obviously not .odt documents. Consequently what can be flagged in ODF XML is not the same. In particular the options in Tools>AutoCorrect>AutoCorrect Options, Word Completion tab cannot be saved: Collect words or not and When closing a document, remove the words collected from it from the list.

I guess that some “normalization”/parse step occurs when loading a .docx and converting to internal representation.

Does the same occurs when you handle .odt documents?

If no, then work with .odt native documents and convert them to .docx only in the final stage when you transmit a finalised document to some external recipient.

Thank you for replying. Sorry, I should have written that the same thing happens with both .docx and .odt documents. I’ll update my first post.

What seems odd is that it’s not possible (or not for me anyway) to delete any collected words from any document. They’re removed from the list, but the moment I close the document and reopen it, or any other document, the item(s) I deleted all reappear.

This is true for both LibreOffice 7 and previously with LibreOffice 6.

Are you able to delete a collected word, and it stays removed?

From a quick experiment, it looks like impossible to empty permanently the list. What I tried:

  • type a short text and check the list: words are created in the list
  • remove a word from the list
  • save document, without quitting, check the list: OK, removed word is not in the list
  • type a few extra sentences and check the list: removed word is back though it was not typed in

Second experiment:

  • type a sentence
  • remove a word from the list
  • save and quit
  • reopen document and check the list: word is back, whether or not Remove when closing is checked or not

My conclusion is: whenever you modify your document, text is parsed again and words are collected. Consequently, it makes no difference whether the collected list is initially empty or not. When you check AutoCorrect list, it will have been reconstructed and you can’t tell if this is the same as the saved list

The only difference could be in file size but I didn’t check this.

2 Likes

Thank you for taking the time to run tests. I can confirm the repopulated list is both document specific and in each case, whichever words I delete, those same words are the ones added back in on reopening that document.

There seems to be an embedded list for each document, but even extracting the document as a zip file failed to reveal the added words list.

Maybe I need to report it as a possible bug, but it seems odd that, if it was a bug, it’s survived through at least two completely different versions of LibreOffice and the Help section guide to deleting Collected Words, must have worked at some point.

As a further quick test, I copied the text from a document, pasted into notepad. Saved that, then copied that unformatted text and pasted it into a new .odt and .docx document. The Collected Words list, specific to the original document, appeared!

This suggests there is a list of collected words, independent of the document itself. Connected to the words ‘in’ the document rather than the document itself.

Thanks once again for your efforts.

The Help page Fine-Tuning the Word Completion for Text Documents seems fairly explicit that words (of stated length or longer) are collected from the current document. The function is to allow word lists to be transferred to a new document

Thank you. Unfortunately, the issue is not the collection of words, but the removal of unwanted items from Collected Words lists. So far it seems impossible to remove unwanted collected items. They won’t delete.

You are absolutely correct that Collected Words lists can be transferred to new documents, and that is the issue. Each new document I type, suggests items from every list, including every mistakenly collected word/item from previous documents. Essentially, errors, in Collected Words lists, are permanent. The only option is to either continually click away every incorrect suggestion or disable Collected Words completely.

If I highlight an item in the Collect Words list and click delete, it will disappear and then reappear the next time that, or any other document, is opened. You simply can’t remove any item in the list.

As you know, Collected Words don’t even have to be words, they can be numbers, alphanumeric expressions, e-mail addresses, literally anything.

Anyway, sorry, I’m going on, which is unfair. If I can’t find a way to delete Collected Words, I might need to file a Bug Report instead. Thank you once again for your comment.

Here is what I find using 7.2.0.4 and 7.1.5.2:

  • All words over 8 letters (default setting) are collected from all currently open documents
  • If I close a document the words from that document remain in Collected Words. That is, unless the box “When closing a document, remove the words collected from it from the list” is ticked when all words from that document are removed from the list.
  • If I close all LibreOffice programs and open a blank document the Collected Words list is empty. If I write a word over 8 letters it populates the list.

I don’t know what purpose you have in mind, but specialist words might be better added to AutoCorrect Replace or to AutoText (which requires F3 to transform a shortcut to the full word).

Thank you for your efforts. The question is how to remove items from the Collected Words list. Once an item appears in the Collected Words list, it can’t be removed. If you untick When closing a document, remove the words collected from it from the list, this has no impact on items already in the list.

As an example, e-mail addresses that start with words get added to the list and so each time I start to type the first few letters of that word it shows the e-mail. All I wish to do is remove such e-mail addresses and any other unwanted items from the Collected Words list.
Help sections in LibreOffice talk about adding items, but the only option to remove them, is by clicking delete, which doesn’t work.
Thank you once again for your efforts, I think I might need to post a bug report.

What is your intended goal from using Collected words?

Thank you for replying. I wish to use Collected Words to speed up typing, nothing more complicated. But my goal, here, is to remove the unwanted collected items that permanently appear in the Collected Words list. The list contains dozens of unused, unwanted items (not purely words) collected over time by numerous documents, in numerous iterations of LibreOffice. The problem is the existing items in those lists, not new items I may add in the future.

Currently, I can’t clear any of those existing items. If I install LibreOffice from fresh on a new computer with a new Profile 4 folder and no reference to any previous installation, I start with an empty list. However, the second I open any of those previous documents, they will repopulate the Collected Words list with their previously collected words and items. All I wish to do is remove the unwanted entries from those lists.

It would seem that each .docx and .odt document carries (embedded) its own specific Collected Words list, but on analyzing each file, I couldn’t find them. I also couldn’t find where LibreOffice stores the combined collected words lists. Deleting the files in any of the autocorr folders (in both program and Profile 4 folders) has no impact. Deleting my Profile 4 folder in its entirety has no impact, as the moment a previous document is opened, the list gets repopulated.

It’s essentially a reference book with permanent errors that you can’t edit!

Thank you for taking the time to reply, and I apologize if my explanations are being misunderstood.

Writer reads all the documents you have open and compiles the collected words list on the fly from those documents.

On opening LibreOffice with a blank document the collected words list is empty so there is no permanent store of collected words.

If you tick the boxWhen closing a document, remove the words collected from it from the list” the words from that document will be removed from the collected words list the moment you close that document.

If opening other (reference) documents at the time you are writing is causing problems, can you open the reference documents in another instance of LO, e.g. portable LO or a separate installation, or even pdf? I would have thought that a reference document is likely to contain similar words so should not present a problem however.

Thank you for the clarification, it makes a great deal of sense.

I opened a blank document and ensured Collected Words and When closing a document, remove the words collected from it… were both ticked, the Collected Words list was empty, and I closed Writer.

I then opened an existing document, with its own populated Collected Words list, and ensured the same two options were ticked and again closed Writer.

On opening, a blank document, its Collected Words list was empty. However, as I typed, Writer automatically suggested entries from the Collected Words list of the previous document. The only document open was the one I was working on. The words/items being suggested were specific to the Collected Words list of the previous document, not just spell checks or my own dictionary. Even after deleting the original document, Writer was still suggesting items from that list.

Collected Words lists are persistent across all documents, regardless of selecting the option to remove them from the list. The only way to clear the ‘internal’ list is to reboot the computer.

But, the moment a document is opened, the Collected Words list from that document is always added to an ‘internal’ list regardless of telling Writer not to do that.

It also begs the question as to what function the Delete button plays on the Collected Words panel. If Writer is collecting words on-the-fly, the Delete button could only work for text not actually in the document. On the Help page under Collected Words it states, To remove a word from the Word Completion list, select the word, and then click Delete Entry. This doesn’t work, and from the way you explain how Writer collects words, couldn’t work. The user has no control over ‘what’ items are collected, only over the minimum word length and the maximum number of items.

I’m not sure what to do. It would seem to be a bug, even though it behaved the same way with LibreOffice 6. What do you think?

As this behaviour is causing user confusion, file a bug report.

This will make developers to have a look at it and at least clarify the description.

However be aware the bug affects a feature probably considered minor and thus will not be assigned top priority.

You are right. It saves the word list of previous documents in memory (although list shows empty) until LO is closed despite ticking When closing a document, remove the words collected from it from the list.

You could add to bug tdf#67826 . There are two competing but opposite enhancement requests to retain persistence between sessions of LO, tdf#86872 and tdf#118309

Here is a tip that you might find helpful:
You can scroll through the word completion tips using Ctrl+Tab meaning that if I type con and get a tip of consents I can press Ctrl+Tab and change the tip to get constant which is the next word in the list, if I don’t like that I can press Ctrl+Tab again to get controlled

Thank you for running the tests. I’ve filed a Bug report tdf#145096 and included tdf#67826 in the notes of my bug report. Let’s see what happens. Thanks again for your efforts.

@ajlittoz As an update, I have now filed a bug report, 145096. Only time will tell! Thank you for your efforts.

I also wanted to clear the word list. One can delete them one-at-a-time, which is not practical for 1000 words. But one can reduce the max entries to 50, and that will clear most of them.