#135 Customizable Duplicate Removal

[Anonymous]

Hi there,

first of all: An amazing piece of software! It's currently
proving to be *very* useful to me.

However, there is something which I would find really
helpful and maybe some others could also use it:

Problem:
I have imported large sets of references which stem from
different sources. The "exact duplicates" method gets
rid of the very obvious dupes, but many of them still
remain, although sometimes the difference is just a
different key or different linebreaks in the abstract. Thus,
manually sorting them out is really tedious as most of
the time the manual dupe search shows identical
summaries.

Suggestion:
My suggestion is thus to make the aumatic duplicate
deletion a bit more flexible by providing options for
ignoring differences in Key, Abstract, Note, Editor...
(basically all of the less important fields).
This feature would help me a lot!

Cheers to you and thanks again for an already great
program.

jhh

[Adrian]

Hi all,

I am also an avid user of Jabref. Great software!
I fully agree on the need to have a flexible duplicate finding system. In my work I collect entries in various different ways. It can be direct input, copy pasting a bibtex entry from elsewhere, importing from different formats (ris, zotero) and getting a bibtex file from someone else. It is easy to end up with duplicates and too much work to check possibilities one by one.

It would be great to have the possibility to:
-launch a search of duplicates based on customizable criteria (eg part of title, partial author name overlap, name, etc. Ideally a set of criteria should be saved as a preference)
-have the option to warn about possible duplicates at import or addition of new entries (based again on criteria that can be checked)

It is alright if one gets false detection, if final confirmation can be done by the user.

One interesting feature would be that when one entry is removed based on confirmation of being the duplicate of another, all group assignments are moved to the remaining entry.

Another very useful feature would be that if too entries are identified as possible duplicates, then a "Not duplicates" command would prevent the system from showing them as duplicates again before any change is done to any of the entries.

Developers, please consider this. Many thanks!

Adrian