Apache OpenOffice (AOO) Bugzilla – Issue 18169
Second Bib entry with same identifier is ignored
Last modified: 2013-02-07 22:34:54 UTC
I have an extensive database of medline entries (located in a csv text file that is registered as a data source), of which some entries end up having the same identifier (e.g. Blobel1975). If I have two entries that have the same identifier, no matter which one of the two I choose, the first one get inserted into the document. even if I switch the two entries in the insterting dialogue (or chose to edit the entrie in the same diaologue) I always get the first of the two entries and the second one is ignored. Juergen OO1.1 rc2 on Mandrake Linux 9.1 using KDE
I have just tested this with the stadard bibliographic database and and inserted a record with a duplicate shortname, and as you say the shortname list shows both entries but only the first record is inserted even when the second shirtname on the list is selected. My guess is that the solution is to enforce shortname is unique rule in the database. It is against the rules / practice of bibliographic citation for the identifier to be the same. Most bibliographic styles have a method for separating otherwise unique ids eg [wilson:2000a] wilson:2000b]. You would have the same problem in LATEX - the bibliographic reference only holds the identifier, and the bibliographic table generator picks up the id and finds the record and format the table entry. If duplicate ids where in the list of reference the same thing would happen ? the first match would be picked. What we need to do ? 1.explain the need for unique short names in the documentation 2.change the database to ensure unique shortname ids. What you might do to get around this problem, - ensure that your ids are unique Ideas on how to do this - 1.load them into a spreadsheet, sort on shortname order, add a column with a formula that shows a red box where there are duplicate shortnames. Fix manually. 2.Also there are tidy programs that fix up many bibliographic consistency problems in large citation lists in bibTeX files ? they may do unique ids. (you will have to convert to bibtext and back) I think bibliographic programs such as sixpack (linux - but will run on windows if you install the right libraries) can import cvs, fix ids (I think it checks unique and makes new ones if the field is empty) export cvs. You may have to fix up record order of OpenOffice.
Hi! The way you describe it it all makes sense. Nevertheless, I guess for the next version it would be nice, if I could do all that in OO. I'am aware thet the bib module will undergo a complete redesign for OO 2.0 but the bib module as it is right know is merly an interface to a database and not mich more. It would be nice if OO would realize something being wrong and offer the possibility to fix it. Well wait and see for OO 2.o I guess :-) Juergen
reassign to owner of selected subcomponent
Now that the configuration is corrected: reassign to owner of selected subcomponent
ES->BH: This is an enhancement for which we have to decide how to force the input of unique identifiers.
My suggestion is to simply create a unique index for the shortname field in the database. Also if you do this change to the database could you also increase the field sizes as in issue 16268 ? I still can not see that there is any good reason not to do this - It cripples the bibliography not to be able to store real book titles, notes, abstracts urls etc - longer than 75 characters.
To grep the issues easier via "requirements" I put the issues currently lying on my owner to the owner "requirements".