Scanned books marked as serials aren't included in Open Library

This may be the same + quoting problem.

Let's see what happens after the reindex, to get the author text names instead of the tdb names. See also https://bugs.launchpad.net/openlibrary/+bug/126816 second item which describes assigning unique tokens to facet labels, to prevent terms in the facet labels from garbling up the nested searches. Alexis and I just discovered that newegg.com seems to do something similar (using numbers instead of alphabetic strings) for its faceting so that sort of validates the idea.

With the new index, the (douglas lurton) search finds 4 books.

Reopen: I think at least two of these books have been scanned, but they no longer show up as having scans available.

BK brought this up at the books meeting today. I had thought the scanned books were showing up more recently than April but now I'm not absolutely sure. Anyway, archive.org search shows exactly one Douglas Lurton book:

http://www.archive.org/search.php?query=douglas+lurton

finds

http://www.archive.org/details/thepowerofposit00lurtmiss

A similar openlibrary search for "douglas lurton" finds (among others) this book as

http://openlibrary.org/b/OL6071081M

going to the history page for that book and viewing the MARC record shows that it's not the same record as the one in the IA search engine. Examination of the most recent OL Json dump shows that this is the only OL record for the book. Looking at the update log after the json dump doesn't show any new records for the book, but there is some chance that the log is incomplete, so I'm hoping to get a new json dump soon. The current update log still has some malformed records discussed in another bug.

My guess is that we have gotten MARC records for this book from two or more different sources, and deduplication has thrown out all but one of them, which is not the IA one. Any thoughts?

Note: IA search also appears to be missing several of these. Metamanager finds:

http://www.archive.org/details/homebookcomplete00lurtrich
http://www.archive.org/details/yourlifepopularg21lurtrich
http://www.archive.org/details/mymothersbiblesc00lurtrich
http://www.archive.org/details/completehomebook00lurtrich
http://www.archive.org/details/spiritualstrengt00lurtrich
http://www.archive.org/details/makemostofyourli00lurtrich
http://www.archive.org/details/thepowerofposit00lurtmiss

I spoke with Edward online last night about this and he agrees about the deduplication theory.

We should make sure during import/deduplication that scan availability doesn't get separated from catalogue records of scanned books, even if we end up using catalogue data from some other source for the general bibliographic info about that book.

Note, of the above Archive identifiers, thepowerofposit00lurtmiss is the only one found in the IA solr.

Turning this over to Edward since it's a data import issue.

Jonathan Rochkind <email address hidden> gave another example:

http://www.archive.org/search.php?query=title:%22sinners+in+the+hands+of+an+angry+god%22+AND+%28mediatype:texts%29

http://openlibrary.org/search?q=Sinners+in+the+hands+of+an+angry+God&ftokens=mhsncqbxgkup

I must let Jonathan know when this bug is fixed.