Is it possible that the record has a really high skip indicator value in the 245 field which is causing the getTitleSort() routine to return a blank string?

 

- Demian

 

From: Jay Roos [mailto:jayroos@gmail.com]
Sent: Friday, April 26, 2013 11:42 AM
To: Demian Katz
Cc: Mark Triggs; vufind-tech@lists.sourceforge.net
Subject: Re: [VuFind-Tech] Skipped entries for docid....

 

Maybe I'm looking at this wrong somehow? Is the docid reported by alphabrowse indexer the same as the record id used by VuFind? If not, what is it? Is it just a sequence number?

 

This is one of the problem records:

 

Skipped entries for docid 280694 because the number of sort keys didn't match the number of stored values.

Fields in title_sort: []

Fields in title_fullStr: [Thunder Road / James Axler.]

 

However there is no record 280694. The title does exist in VuFind as record 647959. So I'm guessing the other number is from somewhere else.

 

All of that aside, it looks like I have a problem with record going through the title_sort routine.

 

On Fri, Apr 26, 2013 at 9:46 AM, Jay Roos <jayroos@gmail.com> wrote:

I cleared the index and started with just one record and browse indexing works fine. I'm going to pull the marc file into MarcEdit and see if those IDs are in the file somehow.

 

On Fri, Apr 26, 2013 at 9:39 AM, Demian Katz <demian.katz@villanova.edu> wrote:

Deleting the index directory ought to do it!  I’m not sure where else they would be coming from.  It doesn’t hurt to also clear out the spelling/spellShingle directories when resetting the index, but I see no reason why these would have anything to do with alphabrowse indexing behavior.

 

What happens if you run the alphabrowse indexer against a completely empty index?  That might be an interesting test to make sure you’re really clearing out the right directory.

 

- Demian

 

From: Jay Roos [mailto:jayroos@gmail.com]
Sent: Friday, April 26, 2013 10:36 AM
To: Mark Triggs
Cc: Demian Katz; vufind-tech@lists.sourceforge.net


Subject: Re: [VuFind-Tech] Skipped entries for docid....

 

I noticed that a number of the IDs being reported no longer exist in our catalog so I cleared the solr index and did a full reindex to address the issue. Unfortunately, those IDs are still being reported in the index-alphabetic-browse output.

 

To clear the old indexes and rebuild I deleted $VUFIND_HOME/solr/biblio/index/ is there something else I missed where those deleted IDs are hiding?

 

 

On Thu, Apr 25, 2013 at 4:30 PM, Mark Triggs <mark@dishevelled.net> wrote:

Hi there,

I've just built the 1.4 version of browse-indexing.jar with a little
more debugging information.  It should now print the two sets of values
it gets back from the index (whose lengths reportedly don't match).

You should be able to use it by replacing your
'import/browse-indexing.jar' file with this one:

  http://dishevelled.net/browse-indexing.jar

Thanks,

Mark



Demian Katz <demian.katz@villanova.edu> writes:

> Strange, I don't see any count mismatches there (though the slash seems to be throwing off the generation of the sort key -- but that's probably unrelated).
>
> Does this title show up in the browse list, or is it absent?
>
> - Demian
> ________________________________
> From: Jay Roos [jayroos@gmail.com]
> Sent: Thursday, April 25, 2013 12:30 PM
> To: Demian Katz
> Cc: vufind-tech@lists.sourceforge.net
> Subject: Re: [VuFind-Tech] Skipped entries for docid....
>
> Sorry, disregard the previous message. That was actually from a title that had a problem with dewey_browse. Here's what's populated for dewey in that record:
>
> <arr name="dewey-full">
> <str>813/.54</str>
> </arr>
> <arr name="dewey-hundreds">
> <str>800 - Literature</str>
> </arr>
> <arr name="dewey-ones">
> <str>813 - American fiction in English</str>
> </arr>
> <str name="dewey-raw">813/.54</str>
> <str name="dewey-sort">813.00000000</str>
> <arr name="dewey-tens">
> <str>810 - American literature in English</str>
> </arr>

--
Mark Triggs
<mark@dishevelled.net>



 

--
Jay Roos
Information Technology Coordinator
Great River Regional Library



 

--
Jay Roos
Information Technology Coordinator
Great River Regional Library



 

--
Jay Roos
Information Technology Coordinator
Great River Regional Library