I have a set of documents that have a somewhat unusual structure, sample below. The exact nature of the contents is not important, but the gist is that the resulting document has an enormous number of paragraphs, because each line is terminated by an Enter.

The docs vary in size, but even the largest is not out of bounds - around 4 MB. However, Word has a tough time when the docs get that large - unbelievably slow, and prone to hanging completely. Other, 'normal' docs of similar size Word handles with ease, so it's not simply the volume that causes problems. I suspect it's the number of paragraphs, since each paragraph has its own set of properties. Word is not optimized for such a structure - a few hundred paragraphs are child's play. However, a few hundred THOUSAND paragraphs seemingly are not (this particular document has almost 3/4 million such 'paragraphs').

Does anyone have experience with such constructs? Is there a way to convince Word to deal with it, or should I simply go to a different editor? I've tried a few other, text-only editors, like Notepad++ and PsPad. Both open even such monsters immediately, search through the entire text instantly and save changes with no noticeable pause. However, Word has well-known advantages - autosave, built-in VBA, etc., so I am somewhat reluctant to just give it up. But I need to work with these things, and if Word is simply not the proper tool, I'll need to switch.

Nt, pe
.JZ
Ger
.GE


._HS Kynzvart - 29414

.AU
Urkunden
.TI
Die ältesten Urkunden...

.IM
S.l.: s.t., #s.a., ###4.


.RO

.PO

Perhaps there is something wrong with your system.

I just created a document containing 375 pages (7306 paragraphs) of your sample text, then did a Find/Replace with:
Find = ^p^p
Replace = ^p
The process took about 1 second to complete. Even creating said document was a breeze.

As designed, Word can handle documents up to 512Mb in size, including 32Mb of text.

__________________
Cheers,
Paul Edstein
[Fmr MS MVP - Word]

I doubt if it is the system. The exact same behavior manifests in all computers that I have tried these docs, and they are all modern machines with adequate RAM, new OS and new Office. Also, there seems to be some threshhold for this behavior. The size you cite in your test is well within the bounds of what still works quite well. If you want a test of the sizes I am dealing with, duplicate the text to get 15,000 pages, instead of 375.

What you're seeing is more likely to be result of the number of pages you're creating and the repagination overheads that imposes. Try working in draft layout view and disabling background pagination.

Nonetheless, you cannot reasonably expect Word to work as quickly with such lengthy documents as with short ones. And, if you think Word is slow, perhaps you should compare its performance with other word-processing & text-editing programs (e.g. Notepad, Wordpad)...

__________________
Cheers,
Paul Edstein
[Fmr MS MVP - Word]

My apologies, but did you read my post, or just the title? What you have mentioned, both times, is stuff I already addressed in my original post.

1. I do not have 7,000 paragraphs, as in your example, but over 700,000 paragraphs.

2. Word handles documents of this size (cca 4MB) quite easily, if they are in a more normal layout. What -seems- to me is the difficulty with these docs is that each line is a separate paragraph.

3. I already HAVE tried other text editors, including specifically one of the ones you mention, Notepad++. The same text that Word chokes on, Notepad++ opens instantly, finds instantly, and resaves instantly, despite the file size being about twice the size of the Word format (likely due to Word's compression of the .docx format). So does PsPad. I did not try WordPad.

Background pagination is already turned off, and I did try various views - no difference.

I also ran the test on a much larger document - 1458 pages (~500,000 paragraphs) of your sample text, then did a Find/Replace with:
Find = ^p^p
Replace = ^p
153,600 replacements were made.

In Print Layout view (and Draft view), there are most definitely things you can't do while Word is repaginating the document after such a Find/Replace. Hence, it will be slower when repagination is allowed. It will also be slower if Word is continually trying to spell-check the document. Turning off auto-save would likely also increase responsiveness in such a large document.

Word is a word-processor, with all that entails. Clearly, you're not using Word for the purpose for which it was designed.

Word also handles ordinary 4Mb word-processing documents in its stride. I have some as large as 15Mb and more, all of which Word handles with ease.

Since you already have a solution (Notepad++) for your peculiar requirements, use that.

__________________
Cheers,
Paul Edstein
[Fmr MS MVP - Word]

And how well did it perform? And did you try it with the 15,000 pages, 750,000 paragraphs I have specified? Again, there seems to be a threshhold up to which it works tolerably, then suddenly degrades dramatically. I could try experimenting with various sizes, but it's slow, so I posted my question to see if anyone has run into this, or has found some documentation on the matter, and maybe already knows what I want to know, without me (or them) having to conduct a laborious and tedious series of tests.


Pagination is already turned off, as I wrote. None of the various views makes any difference that I can discern. Spell-check might help, but that's a pain in the neck to constantly turn on and off, depending on which document I have open, but if that would make difference between usable and not, I could live with it. Doesn't seem to help, though, or at least not enough to notice.


I don't know about that - this is text, and I thought it should handle text, regardless of layout. I -think- it has something to do with the unusually large number of paragraphs, but I'm not sure - it's just my impression, hence my question, whether anyone knows of such limits or restrictions.


Yes, I know that. I wrote that myself, that I have documents of similar size but more conventional layout, all of which work normally. Word handles them without the slightest hiccup.


I have several possible avenues, Notepad++ is one of them. But again, I'm looking for some definite information on the subject, before I abandon Word altogether.

Direct formatting can have a tremendous clogging effect on Word in large documents.
Importance of Styles in Word
Take a look at the experiment I ran on the page linked above. (It is shown in a box.)

See this thread. Computer Slow with 5 files https://www.msofficeforums.com/word/43721-computer-slow-5-files.html

Can you explain what it is you actually do with the content of such a large document? It doesn't sound like Word is the best tool for this scale of content because I'm not sure how you would ever totally avoid the WYSIWYG tendencies that underpin Word. It isn't just holding the text, it is continuously working out where that text sits on a page based on the formatting applied to it. Notepad++ (or a database for that matter) doesn't have the same processing overheads.

If you are considering staying with Word, the suggestions I would have are much like those of Charles and Paul above:
1. Minimise repagination by using either Draft (or possibly Web view)
2. Remove all local/direct formatting by selecting all and pressing Ctrl-Q and Ctrl-Space
3. Turn off all proofing so background spell/grammar is not happening

I'm pretty sure the drastic slowdowns you are seeing are unavoidable in Word so I think if it were me, I'd be looking to other software packages rather than persisting with Word.

__________________
Andrew Lockton
Chrysalis Design, Melbourne Australia

Removing all formatting did help, a little, but it was still slow as molasses to respond. I then wrote a short VBA routine that changes 99 Enters to spaces and leaves only the one-hundredth one as a paragraph ending. Let it run all night, and the paragraph count went down from over 730,000 to 690,000. I'm not interested in waiting several weeks for it to finish.

I then used find/replace to manually change most of the Enters to spaces, using accompanying patterns, so I wouldn't change the entire doc to one long paragraph. That got the count down to 24,000 and pages down to 1400. A close and reopen to flush the toilet, and the new doc works normally, despite the number of character and words staying the same - just the paragraph and page count are more normal. As I was doing the find/replace, the response time got steadily better, and after close/reopen became completely normal. I didn't run any actual timing tests, so it's pretty subjective, but even subjectively, the difference between minute-long lags/hangs and instant response is clear enough.

Removing formatting helped some, but not enough. Disabling spell-check had no effect, which is understandable - word are spelled or misspelled the same way, regardless of how many paragraphs they are in.

I suppose I could reverse the edits and progressively go back to the quarter million paragraphs, running timing tests along the way, but that sounds like a lot of work for not much useful information. It's now clear to me that for a document structured this way, Word is simply not the proper tool for the job.

Yes, I've come to that same conclusion. I've decided to go with Notepad++ - I already tested it, and it does everything I need, with instaneous response times. It also has automatic formatting by content, which Word does not - you have to apply formatting to the actual text, either by styles or directly, but you have to do it. Notepad++ allows a formatting definition, so specific patterns of text are automatically displayed formatted. That also serves as a check on content. If I define, for instance, the pattern .AU to be formatted in blue and bold, and the user accidentally types .AV, the formatting does not apply and they have an immediate visual clue that they made a mistake.

What the users actually need to do with the contents is pretty much the same as any other text document: search/replace, and manually edit the text. I know it looks weird - it's a legacy format that serves as the source for a historical book and manuscript library catalogging system. The results then get vacuumed up by Access into a custom database, but that's outside the scope of this issue, and it works well in any case.

Word is great for a great many things, and is the default for any work with text in our institution, but there are peculiar cases that give it fits, and this is clearly one of them.