Index status not clear

danny layman
I cleared my index, save webpages, and tried to rebuild it. Search yield quite a few results and it seems there is no real problem, all is indexed. But the prefs dialog report 6 indexed files and 531 unindexed ones. Database integrity check is ok too. What is going on? (win xp, ff 9, zotero 3.0.1) And is there a typical words number for this amount of files?
  • dstillman
    Is that still the case after restarting Firefox?
  • danny layman
    Yep, and after a pc restart too. Now I have noticed some other sympthoms:

    The sqlite file is significantly smaller than the sqlite.bak from about two days ago. (How would I restore the bak?)

    Export Library procedure crashes in a certain point, leaving a 197 KB .rdf and a files folder with 2400~ files. The Zotero storage folder contains around 18,000 files.
  • dstillman
    Are you sure it's searching full-text content, then?
  • dstillman
    http://www.zotero.org/support/zotero_data#restoring_your_zotero_library_from_a_backup
  • danny layman
    Oh, I have long forgotten the search field is flexible, and now i understand search results really are defected. Anyway, it is set as default to index 500K characters, 100 pages.

    Now one reason I don't really want to reinstall and restore the lib is that I use Date Added to sort it all. When I used the folder/rdf method once the dates were overwritten. Will they be saved if restored from the site?
  • dstillman
    An RDF file isn't a backup.
  • adamsmith
    and restoring from a back-up will keep date-added intact.
  • dstillman
    But if you want, before you restore from a backup we can try to figure out why it's not indexing anything. Obviously it should.

    You can see the index state of PDFs in the right-hand pane. Does indexing individual PDFs work?
  • danny layman
    I do. It does
  • dstillman
    And what about after you run Rebuild Index? Do individual PDFs show as indexed?
  • danny layman
    Let's continue. They do.
  • dstillman
    But you're saying you can't find text from a PDF that shows up as indexed using the "Attachment Content" Advanced Search condition?

    If not, upload your database to the DB Repair Tool and post the Upload ID from the tool, and we can take a look. You can ignore the download link the tool provides.
  • danny layman
    Is that a double negation? I can.
  • dstillman
    edited February 12, 2012
    No idea what you're talking about. You seem to be saying that after rebuilding your index, the items show as indexed, and also that you can find text from the items that show as indexed. That sounds like everything is working to me. If it's not, you'll have to respond a bit more verbosely, or else we're just wasting each other's time.
  • danny layman
    Sorry. The PDFs I index manually become searchable and add to the Indexed count. A wholesale rebuild still doesn't fetch the rest of them.

    Should I upload it then?
  • dstillman
    Yes, you should.
  • danny layman
    That's ID 4f383b25bb6c9
  • dstillman
    edited February 12, 2012
    Can you clear your index, restart Firefox, and provide a Debug ID for a reindex attempt (from the prefs)?
  • danny layman
    The Debug ID is D421089774.
  • dstillman
    edited February 13, 2012
    The issue here seems to be bad Unicode data in one of your files. It's failing because of a Mozilla bug, but I've put in a workaround that should fix at least this case. You can try it in the 3.0 Branch dev XPI or wait for 3.0.2 later today.

    (There's another case where this could fail, and there's a chance you'll hit that too if you have other files with bad data. Let us know.)

    You should also upgrade to Firefox 10. You're still on 8.
Sign In or Register to comment.