To flesh out this answer, Shotwell has two mechanisms for detecting duplicates: the MD5 checksum of the thumbnail and an MD5 checksum of the full file.
The issue that's mentioned in the message Mattias linked to essentially means Shotwell will use only a full MD5 checksum to detect duplicates: http://trac.yorba.org/ticket/2587 Shotwell has never used a filename to detect duplicates. My guess is your duplicate files either have duplicate thumbnails (possible) or really are byte-for-byte duplicates. For that, there's another feature request you should be aware of: http://trac.yorba.org/ticket/2593 -- Jim On Tue, Nov 9, 2010 at 5:49 AM, Mattias Põldaru <[email protected]> wrote: > Ühel kenal päeval, T, 2010-11-09 kell 12:40, kirjutas Richard Booth: > > Hi > > > > I have around 8,000 photos that I took years ago that I want to import > > to Shotwell. When I tried importing Shotwell said there were 1500 > > duplicates. However I think this is because my old digital camera > > reset the file numbering after each batch, thus I have a number of > > DSCF0001.jpg etc. > > > > I'm thinking I could write a script to append a serial number to my > > 8,000 photos and then import to Shotwell. Will this work? Am I right > > in assuming Shotwell says it is a duplicate based on the filename > > alone? > > No, it compares MD5sums of EXIF thumbnails (maybe more). See this answer > from Jim: > http://lists.yorba.org/pipermail/shotwell/2010-November/001199.html > > > Regards > Mattias > > _______________________________________________ > Shotwell mailing list > [email protected] > http://lists.yorba.org/cgi-bin/mailman/listinfo/shotwell > _______________________________________________ Shotwell mailing list [email protected] http://lists.yorba.org/cgi-bin/mailman/listinfo/shotwell
