Finding elements after Scriv3 project conversion

After I converted a Win Scriv1 project to Win Scriv3 format, where are my collections and keywords hiding?
(For the collections, the Collections menu in the sidebar is empty; for the keywords, the Keywords popup is empty.)

Win 10 x64 [19042]; i7 7700, 32GB, 2T.
Scrivener 3.01.

[Conversion lost collections? Conversion lost keywords?]
[‘Edit’ was to add ‘notify’.]

I understand that there is rush for support on the Windows Scrivener v3 – Congratulations :wink:
However, it has been a couple of weeks since I asked this upgrade question; I’ve yet gotten any guidance. :unamused:

Having gone through the project conversion twice with the same result – lost Collections and Keywords – I’ve been forced to continue using my ‘old’ v1.9.16 … :frowning:
(The conversions for the 447 MB project took roughly 30 minutes, as I recall from those two weeks back.)

To repeat that first post:
“After I convert a Win Scriv1 project to Win Scriv3 format, where are my collections and keywords hiding?
(… the Collections menu in the sidebar is empty; … the Keywords popup is empty.)”

Win 10 x64 [19042; i7 7700, 32GB, 2T.
Scrivener 3.01.

Our upgrade guide is here: … -guide-win

Does that answer your questions?


Thank you Katherine,
Re the the upgrade guide’s answering: No, not that I’ve noticed. For instance, SEARCHing for “collection” in the guide yields only two hits, one being the relevant menu item (and the related key shortcut) and one being for the “Freeform corkboard”;
for “Keyword”, no hits.

After I converted the project (twice), I had both times found my way to the View menu’s Collection item, which already showed Collections checked; and I did similarly get to the Project menu’s Project Keywords item, which popped up the project’s Keywords box, that, empty of the project’s 16 keywords.
Re Collections, the sidebar does not show any of my 41 collections below its Collections bar, and mousing over the bar’s lower edge does not produce the usual double line/arrow to expand or reduce its view so apparently nothing’s hiding. My adding sufficient new collections demonstrated that the view is working as expected. (I see that the max collections visible is still about seven: It would be of use were I able to see as many as the window would allow.)

What am I missing?

This sounds to me more like a bug of some sort than anything intentional. There should be no need to really do anything to have your keywords and collections all showing up right where you would expect them to be (visuals aside, not much has changed with the mechanics of any of that).

In my testing, I don’t see any problems upgrading these two elements fully, with all assignments carrying over in both cases. You might want to send this project in to technical support so that it can be reproduced and fixed!

Thanks Amber, I’ll swing over to support and send it along. g

Amber et al.,

Sending you folks the project is not feasible: My upload speed is far too slow (~5 Mbps) and the project is a rather large 500 MBs.
I will pare it by steps as time permits, ultimately to something about a tenth, testing conversions along the way.

(In another thread, there was a concern about the number of its document files, ~ 25k, though I’ve had no routine issue with that – particularly since I upped Scrivener’s priority!)

Understood! That’s a large project to transfer even if you have a good connection. There are ways to send a sample project that is trimmed down, without too much work on your part. It would result in a project that is functionally useless to you, but for what we would need to be testing here, it would likely be good enough (mainly we’re concerned with the keywords metadata and collection assignments, which have nothing to do with data).

  1. First, create a full copy of the v1 project that you intend to dismantle, using File Explorer to duplicate the master “.scriv” folder. Make sure it has a name that you won’t ever confuse with the real deal, like “My Original Project Name DEBUGGING.scriv”.
  2. In Explorer, go into the Files\Docs subfolder. You should see bunches of numbered RTF files, maybe some TXT and research documents like PDF.
  3. Just trash everything in that Docs folder.
  4. Lastly return to the top level of the project folder, and delete the Snapshots folder if necessary.

The total project size should now be extremely minimal, only the architecture and metadata of the project, along with your binder entries. If you were to load the project in v1, you would see every binder item empty. Again though, that’s fine because this isn’t a bug in the data upgrade but in the structure of the project itself. So that big difference aside, make sure everything else looks like it should. You can skip previewing this in v1 if you no longer have it installed.

Verify the bug still persists by upgrading it with v3. If it does, then zip up the original “Backup” copy and send it in.

Thanks for any help you can provide! Sorry to put you out, but if it’s a problem we can reproduce and solve, it will potentially help other people as well, and we may be able to offer advice to getting the main project upgraded properly once we understand the issue.


Thanks Amber,
The instructions got me to a skeleton that still failed the conversion to WScriv3 much more quickly than my gradual reduction would have.

The zipped ~ 10 MB file is on its way to support.

Many thanks for taking the time to do that! Hopefully we can get that fixed up for the next update.

Oh, come to think of it, 10mb may be over the limit of what our help desk accepts as attachments. You might want to use a Dropbox (or similar) share link instead. That’ll probably take less time uploading for you as well, since it may already be uploaded.

Thanks Amber; I’ve sent a download link to support; it’s for a site called Jumpshare.
The link is good for seven days. (I forgot to mention that in the resent email to support.)

The 10.6 MB file took very little time to upload.