Replacing characters in a large file does not work as intended

intrelis · September 5, 2021, 1:16pm

I have a pretty large document, 1500 pages, and I am trying to remove some particular characters in the whole file and am having a very strange issue.

Say %, LO finds 150 instances and replaces them, but when I scroll down, the same symbol is still in the text in multiple places. I can select it, copy, cut, delete manually, but if I try to use find/replace again, LO does not recognize it. Visually it is there, I can manipulate it manually, but as far as automatic search and replace is concerned the program acts as if it has already been replaced.

Any ideas what is happening and how to fix it? Much appreciated.

EDIT: Windows 10 pro (x64), LO 7.1.2.2

LeroyG · September 5, 2021, 1:45pm

Have you tried to copy one of those rebell symbols and paste it in the Find: field?

When you put the cursor at the right side of a % symbol and press Alt+X, do you see U+0025?

Can you share your operating system and LibreOffice version? To do that, choose menu Help - About LibreOffice, and copy from there.

Edit your question (the pencil icon) to add the new information. Thanks.

Wanderer · September 7, 2021, 6:29am

Can you copy the special character or a paragraph with them in a new writer document (manually of course) and upload here?

Otherwise check, if special options are selected for find and replace.

intrelis · September 8, 2021, 9:03pm

Sorry for the late reply.
Yes, I see the UTF code and I can also copy/paste the character in the find box and the current situation is that after replacing it X number of times, if I click find, it finds the next one (which was supposedly replaced) and then I can “replace all” again for the same X amount of instances. I have no idea what is going on.
Edit: If I let the file stew for a while (without doing anything), it once again stops being able to find this character, but it is still there.

intrelis · September 8, 2021, 9:12pm

I don’t think the characters themselves matter, I have tried with anything and the result is the same. But I have uploaded a page to see if there is anything about the text/file that is not right. The characters I’m trying to replace (remove) are the superscript ones and the odd quotes and such (this is an OCR file, so it requires proofing), like the ones listed at the bottom.
By the way, this issue persists in the new, 1 page file.
ReplaceChars.odt (17.6 KB)

mikekaganski · September 8, 2021, 9:48pm

In the file you attached, I try to select characters listed at the bottom, one by one; after selecting one of them, I press Ctrl+H, and then Replace All. Then I check using Find All.

I have repeated the procedure for all the characters there, and I don’t see any remnants. Using 7.2.1.1.

ReplaceChars-done.odt (16.0 KB)

LeroyG · September 9, 2021, 12:39am

In the Find: field I wrote \^|«|“|»|°|", with Regular expressions, and all was OK.
But could’t find the superscript characters. Need to try a bit more (but not today, …or yes).
To find the а|у characters, you need to select the format automatic superscript.
By the way the superscript а (U+0430) is not a regular a (U+0061).
I guess that the lag in replacing may be due to the repaging process.
LO 7.0.6.2, Windows 10.0.

intrelis · September 9, 2021, 11:27am

Yes, I am aware of how to find the characters, that is not my issue, as I have explained above. I would be fine with lag, if it meant it eventually got done, but that’s not the case.

The a is not a regular Latin small a (U+0061), because it is a Cyrillic a (U+0430), as is the entire text. That should not be an issue, though, never had problems with Cyrillic in LO.

intrelis · September 9, 2021, 11:30am

@mikekaganski Interesting, would that mean it is something particular with my system? How should I go about troubleshooting it? I see that my version is slightly older, I will update it and try again.

I downloaded OpenOffice and it worked fine with replacing any and all characters, but it took 5-10 minutes every time to open/close/save the file and I really can’t deal with that.

mikekaganski · September 9, 2021, 11:34am

Yes, please try that first. Actually, I don’t have other suggestions (other than in case the update doesn’t help, try running safe mode/resetting user profile).

Note that 7.1 vs 7.2 is a major update (LibreOffice version numbering treats two first numbers equally important).

intrelis · September 9, 2021, 4:51pm

No luck. The issue persists.
When I “replace all” any character I see the ones on my screen get removed. If I try to find it again, it is gone, but if I replace another character in the meantime and then go back to the first one, I can “replace all” again and so on on repeat. I should probably contact LO support or something, this is way over my head.