Hi Guido:
On Wed, 22 Jul 2009, Guido Pelzer wrote:
> i have problems with the search results of your search engine (API)
>
> http://www.digitalisiertedrucke.de/search?d1=2009-07-13%2000:00:00&d2=2009-07-22%2000:00:00&dt=c&of=xm
>
> for the period 2009-07-13 - 2009-07-22 you get 9837 records - that's correct
>
> BUT
>
> http://www.digitalisiertedrucke.de/search?d1=2009-07-21%2000:00:00&d2=2009-07-22%2000:00:00&dt=c&of=xm
>
> for the period 2009-07-21 - 2009-07-22 you get 490805 records - that's
> all records of the database. it's wrong because i haven't bibupload
> any record.
This is because there were no hits, and the date limit is considered as
a kind of `additional' limit only, and the search engine's default in
case of no hits with limits is to discard these additional limits in
order to show the `main' results to the user. If you open the second
URL without the `of=xm' bit, then you will see a warning message:
No match within your time limits, discarding this condition...
So, if you want search engine to produce exact search results only, then
you can append `ap=0' (=alternative patterns switched off) to the URL:
<http://www.digitalisiertedrucke.de/search?d1=2009-07-21%2000:00:00&d2=2009-07-22%2000:00:00&dt=c&ap=0>
which will return zero hits.
Please see the documentation for the `ap' parameter at:
<http://invenio-demo.cern.ch/help/hacking/search-engine-api>
P.S. As I started to do since last week, I'm CC-ing the developers list,
when the topic is public and may be of general interest. I hope
that's fine with you, and please don't hesitate to use the list
directly in the future.
Best regards
--
Tibor Simko ** CERN Document Server ** <http://cds.cern.ch/>
