[Refdb-users] Endnote to RIS to RefDB troubles

[Markus H. <mar...@mh...>]

Hi,

Marc Baaden writes:
 > An example ris file that reproduces those can be found at
 > http://baaden.nerim.net/pub/err.ris
 > 
 > Being rather a newby, I wonder what is going wrong and how to
 > correct it ?
 > Does the Endnote RIS export filter need some updating, or
 > should I write a perlscript to tweak some probably malformed
 > entries ?
 > Obviously with 4000 references, you don't want to correct the
 > several hundred errors by hand :)) (To be precise, 3400 dataset(s)
 > added, 352 failed)
 > 

I'll have a look at the ris file. I've experienced a few problems with
the 0.9.2a release myself, all of them were related to bugs in the
auto-creation of citation keys (the bugs are described on the refdb
project page at sourceforge). These bugs have been addressed in the
CVS version. I'll find out whether upgrading would help in your case.

 > 
 > == Keeping IDs ?? ==
 > 
 > As I have documents which reference the bibliographic entries
 > by the actual ID numbers from Endnote, I would like to keep
 > that numbering. I have tried with the -k option like in
 > 
 > refdbc -d bibthese -C "addref -k Biblio_THESE.ris"
 > 
 > But get the following errors continuously:
 > 
 > ID 19 ignored, I have assigned a new one
 > try to add set as reference 293
 > 
 > What am I doing wrong ?
 > 

There is currently no way to convince the SQL database server to reuse
existing ID values (they have to be unique, so it is easier to let the
server create them). The -k switch simply puts the previous value into
the U5 field so it's not lost. You could retrieve references by their
previous ID with something like :U5:=1234. 

 > 
 > == Performance issues ==
 > 
 > The database I am converting is about 4000 references, and
 > it took more than an hour to add those references to refdb.
 > As I am currently experimenting, I might have to do this a
 > number of times until the result is acceptable. Is there a
 > way to speed this up ?
 > 

Could you provide some details about your setup? Which database
server, what computer, do you use RefDB across a network etc? I don't
have any sort of "benchmark" values available, but an hour sounds like
a lot to me. How long would EndNote take to swallow the data?

regards,
Markus

-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de