[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: *SPAM* Re: submission/search problems on demo site
- From: "Stuart L. Marshall" <marshall@xxxxxxxxxxxxxxxxx>
- Subject: Re: *SPAM* Re: submission/search problems on demo site
- Date: Fri, 22 Jul 2005 09:50:58 -0700
Quoting Tibor Simko (tibor.simko@xxxxxxx) on Fri, Jul 22, 2005 at 05:42:32PM +0200:
> Hello
>
> On Fri, 22 Jul 2005, Stuart L. Marshall wrote:
> > It is "000000153 980__ $$aPICTURE"
>
> Okay, so everything seems to be fine on the record side.
>
> Here's what you can try now:
>
> $ cd PREFIX/lib/python/cdsware
> $ python -c "from search_engine import search_pattern; > print search_pattern(p='PICTURE',f='collection').tolist()"
>
> Is the record #153 among the ones found?
Yes:
[1, 2, 3, 4, 5, 6, 7, 77, 78, 79, 80, 81, 82, 83, 153]
>
> 1. If yes, then the collection word index is up to date, and your
> problem may be connected to the collection cache. Try:
>
> $ webcoll -f
>
> to force the collection cache update.
>
!This changed the "Search XXX records for:" from 152 to 153 so that
is good.
> 2. If not, then the collection word index may not be up to date, or
> you may be using the stemmer for BibIndex that may break
> correspondence between the collection definition and the tag 980.
>
> 2a. If you feel the index may not up to date, then check whether
> record #153 is present in the inverted index:
>
> $ echo "SELECT id_bibrec FROM idxWORD02R WHERE id_bibrec=153;" | > PREFIX/bin/dbexec
>
> and if not, try to force the index update:
>
> $ bibindex -wcollection -a -i1-1000
>
> 2b. If you are using stemmer, then change your collection
> definition via WebSubmit Admin to read
>
> collection:"PICTURE"
>
> with quotes included. This will ensure that the search is done
> in the phrase index, not in the word index, where the word
> PICTURE might have been stemmed away into PICTUR, for example.
>
> 3. Independently, does the record appear if you do a regexp search?
>
> $ python -c "from search_engine import search_pattern; > print search_pattern(p='^PICTURE$',f='collection',m='r').tolist()"
Yes:
[1, 2, 3, 4, 5, 6, 7, 77, 78, 79, 80, 81, 82, 83, 153]
>
> Tibor
So now what? I must have done something wrong that prevented the collection
cache from updating.
I'll try changing the bibsched task schedule as you suggested earlier and
then upload another document to see if it becomes visible.
fingers crossed,
Stuart
|