Summary: ~2,000 broken PDF attachments, ~2,900 missing PDFs on disk, metadata and full\-text indexes
### Short summary
My Zotero library appears healthy, but a large number of PDF files have disappeared from the storage directory.
- Zotero version: 10.0.8 (Windows)
- No plugins installed
- Bibliographic records are still present
- Full\-text indexes (.zotero\-ft\-cache) are still present
- Storage folders are still present
- PDF files themselves are missing
Database analysis shows:
- 4,949 PDF attachments referenced by Zotero
- 2,017 PDF files physically present
- Approximately 2,932 PDF files missing
Within Zotero, I can identify about 1,996 items whose PDF attachment appears as a greyed/ghosted PDF icon and cannot be opened.
My main goal is to understand:
1. what could have caused this,
2. whether anyone has seen this before,
3. and what is the safest way to recover as many PDFs as possible.
---
Hello,
I'm trying to understand what happened to my Zotero library and what recovery options are available.
**Environment**
- Zotero 10.0.8 (64\-bit)
- Windows
- No plugins installed
- Zotero file storage quota: 300 MB (full)
- No WebDAV configured
**Symptoms**
Many entries in my library have a PDF attachment that appears as a ghosted/greyed PDF icon.
When I try to open one of those PDFs, Zotero reports:
"The attached file could not be found..."
The bibliographic records are still present, but the PDF files themselves are missing.
**What I found**
I inspected my Zotero data directory and database.
Storage statistics:
- 5,554 storage folders
- 2,017 PDF files physically present
- 5,546 .zotero\-ft\-cache files present
Using SQLite queries on zotero.sqlite, I found:
- 4,949 PDF attachments referenced in the database
- Only 2,017 PDF files actually present in storage
Therefore approximately:
- 2,932 PDF files are missing from disk
**Important observation**
For most missing PDFs, the corresponding storage folder still exists and contains a .zotero\-ft\-cache file with the full extracted text of the article.
Example:
- PDF missing
- .zotero\-ft\-cache still present
- Full article text still available inside the cache file
This suggests that the PDFs definitely existed locally at some point and were successfully indexed by Zotero.
**Things already checked**
- Windows Recycle Bin is empty
- No OneDrive involvement for the Zotero data directory
- No Zotero plugins are installed
- Zotero database appears healthy
- zotero.sqlite, zotero.sqlite.bak, and zotero.sqlite.1.bak are present
- No "Previous Versions" available on the Zotero folder
- No obvious duplicate Zotero data directory elsewhere
**Additional observations**
Using a filtered view, I identified about 1,996 items whose PDF attachment is shown as missing (ghosted icon).
Zotero clearly knows that these attachments are broken, but because the attachment record still exists, I do not see an obvious way to bulk re\-fetch PDFs from DOI metadata.
Many of the affected items still have valid DOI fields.
I also noticed that deleting one of these ghosted PDF attachments is not straightforward because some of these entries appear to be standalone PDF items, and deletion may move the entire item to the trash rather than simply removing a broken attachment.
**What I do not understand**
The structure of the Zotero data directory appears largely intact:
- storage folders remain
- .zotero\-ft\-cache files remain
- HTML snapshots remain
- bibliographic metadata remain
- SQLite database remains healthy
Only a large number of PDF files appear to be missing.
This makes me wonder whether this is:
- a known Zotero issue,
- an external file\-system event,
- a synchronization problem,
- or something else entirely.
**Questions**
1. Is there a known Zotero mechanism that could result in PDF files disappearing while keeping:
- bibliographic records,
- storage folders,
- and .zotero\-ft\-cache files?
2. Has anyone seen a similar situation before?
3. Is there a recommended method to bulk re\-download PDFs for items that:
- already have attachment records,
- but whose attachment files are missing?
4. Is there a safe way to identify and remove only broken PDF attachment records without deleting parent bibliographic items?
5. Is there an existing script or Zotero API approach to:
- detect missing attachment files,
- remove broken attachment records,
- and run "Find Available PDF" in bulk?
Any advice would be greatly appreciated. At this point my main goal is recovering as many of the approximately 2,000 active broken PDF attachments as possible.
Thank you.
My Zotero library appears healthy, but a large number of PDF files have disappeared from the storage directory.
- Zotero version: 10.0.8 (Windows)
- No plugins installed
- Bibliographic records are still present
- Full\-text indexes (.zotero\-ft\-cache) are still present
- Storage folders are still present
- PDF files themselves are missing
Database analysis shows:
- 4,949 PDF attachments referenced by Zotero
- 2,017 PDF files physically present
- Approximately 2,932 PDF files missing
Within Zotero, I can identify about 1,996 items whose PDF attachment appears as a greyed/ghosted PDF icon and cannot be opened.
My main goal is to understand:
1. what could have caused this,
2. whether anyone has seen this before,
3. and what is the safest way to recover as many PDFs as possible.
---
Hello,
I'm trying to understand what happened to my Zotero library and what recovery options are available.
**Environment**
- Zotero 10.0.8 (64\-bit)
- Windows
- No plugins installed
- Zotero file storage quota: 300 MB (full)
- No WebDAV configured
**Symptoms**
Many entries in my library have a PDF attachment that appears as a ghosted/greyed PDF icon.
When I try to open one of those PDFs, Zotero reports:
"The attached file could not be found..."
The bibliographic records are still present, but the PDF files themselves are missing.
**What I found**
I inspected my Zotero data directory and database.
Storage statistics:
- 5,554 storage folders
- 2,017 PDF files physically present
- 5,546 .zotero\-ft\-cache files present
Using SQLite queries on zotero.sqlite, I found:
- 4,949 PDF attachments referenced in the database
- Only 2,017 PDF files actually present in storage
Therefore approximately:
- 2,932 PDF files are missing from disk
**Important observation**
For most missing PDFs, the corresponding storage folder still exists and contains a .zotero\-ft\-cache file with the full extracted text of the article.
Example:
- PDF missing
- .zotero\-ft\-cache still present
- Full article text still available inside the cache file
This suggests that the PDFs definitely existed locally at some point and were successfully indexed by Zotero.
**Things already checked**
- Windows Recycle Bin is empty
- No OneDrive involvement for the Zotero data directory
- No Zotero plugins are installed
- Zotero database appears healthy
- zotero.sqlite, zotero.sqlite.bak, and zotero.sqlite.1.bak are present
- No "Previous Versions" available on the Zotero folder
- No obvious duplicate Zotero data directory elsewhere
**Additional observations**
Using a filtered view, I identified about 1,996 items whose PDF attachment is shown as missing (ghosted icon).
Zotero clearly knows that these attachments are broken, but because the attachment record still exists, I do not see an obvious way to bulk re\-fetch PDFs from DOI metadata.
Many of the affected items still have valid DOI fields.
I also noticed that deleting one of these ghosted PDF attachments is not straightforward because some of these entries appear to be standalone PDF items, and deletion may move the entire item to the trash rather than simply removing a broken attachment.
**What I do not understand**
The structure of the Zotero data directory appears largely intact:
- storage folders remain
- .zotero\-ft\-cache files remain
- HTML snapshots remain
- bibliographic metadata remain
- SQLite database remains healthy
Only a large number of PDF files appear to be missing.
This makes me wonder whether this is:
- a known Zotero issue,
- an external file\-system event,
- a synchronization problem,
- or something else entirely.
**Questions**
1. Is there a known Zotero mechanism that could result in PDF files disappearing while keeping:
- bibliographic records,
- storage folders,
- and .zotero\-ft\-cache files?
2. Has anyone seen a similar situation before?
3. Is there a recommended method to bulk re\-download PDFs for items that:
- already have attachment records,
- but whose attachment files are missing?
4. Is there a safe way to identify and remove only broken PDF attachment records without deleting parent bibliographic items?
5. Is there an existing script or Zotero API approach to:
- detect missing attachment files,
- remove broken attachment records,
- and run "Find Available PDF" in bulk?
Any advice would be greatly appreciated. At this point my main goal is recovering as many of the approximately 2,000 active broken PDF attachments as possible.
Thank you.
Upgrade Storage
If you did have the files on this computer, something deleted the PDFs on disk outside of Zotero, or you're using a different data directory from the one you were using previously and Zotero pulled down your library fresh. (That would be equivalent to being on a different computer.)
Zotero syncs full text separately from attachment files by default, so the presence of .zotero-ft-cache files isn't meaningful in any way.
If you still have the computer where you added the files, you'd have to transfer the complete library.