Samueli Yes, that needs to be re-defined. When a record is deleted, it is 99% of the time a duplicate record that has been found, thus citations on both records "double count". This is pretty important, we get a lot of client questions about citation counts that seem to drop for them.
Deleted records should be unindexed. Mike -----Original Message----- From: Samuele Kaplun [mailto:[email protected]] Sent: Thursday, September 30, 2010 6:43 AM To: project-cdsware-developers Subject: DELETED records definition Dear all, what is the definition of a DELETED record in Invenio? According to BibIndex and BibEdit a record is deleted when "DELETED" has been put tag in 980__c. However, by how the "collection" logical field is defined (i.e. with "980__%"), any subfield of tag 980__ having the value "DELETED" will correspond to a deleted record. Thus BibIndex will remove from its indexes, only record matched by 980__c, while other records will still be indexed, and will need to be manually stripped away (like it happen in webcoll: [...] # B - collection does have dbquery, so compute it: # (note: explicitly remove DELETED records) if CFG_CERN_SITE: reclist = search_pattern(None, self.dbquery + \ ' -collection:"DELETED" -collection:"DUMMY"') else: reclist = search_pattern(None, self.dbquery + ' -collection:"DELETED"') [...] So, records using 980__b or 980__c will be deleted according to WebColl (and thus WebSearch), but not according to BibIndex. They are in a kind of limbo state. Should we enlarge the definition implied by BibIndex? Cheers! Sam -- Samuele Kaplun Invenio Developer ** <http://invenio-software.org/>
