[Digikam-users] Batch delete duplicates?

Johnny yggdrasil at gmx.co.uk
Sun Aug 28 23:24:13 BST 2011


jdd <jdd at dodin.org> writes:

> Le 28/08/2011 18:53, sleepless a écrit :
>
>>> After having used the "find duplicates" tool, I end up with a long list
>
> problem is many photos can have the same name still no be true
> duplicates (edited one versus original ones...)
>
I wouldn't see it as a problem but rather a user decision of how to work
with images that are either edited or nearly identical for other
reasons, e.g. playing with camera settings and sequential shooting
somtimes give very similar images that can be picked up by the duplicate
tool. This will depend on the threshold setting of the "match" criteria,
but it would still be useful to be able to choose e.g. the first or last
version of a photo, or even keep first and last, and skip everything in
between to e.g. enable quick deletion of intermediary edits, revert to
"originals" or keep latest version only. Of course, one could get false
matches depending on the settings used, and what type of pictures the
album/s contains, but that I assume would be a user decision?

An issue for me (maybe others?) in adapting the digital world is that I
like to try to get organised now when the tools are available / I
(well, sometimes at least) understand how to use it, and many legacy
photos I have copied and moved around in a fairly random fashion (a
sporadic backup on a CD, maybe som image on a 3.5" disk still, some on
an old HDD etc.). Some images may be in several places, sometimes with
different names, in particular random caps (upper or lower). I want to
reel these in in a structured way to a common structure (i.e. Digikam),
but will clearly have some duplicates around. I really do not want to
blatantly lowercase all photos (as I now ended up doing), as there is a
risk, albeit small (however, if two cameras would for some reason start
at the same numbering at some point, a significant number of images
would be lost!), that two different cameras have used the same name for
different images which means I would lose images! As I know these images
are sufficiently different not to get false positives, using the
duplicates tools seems to be the right choice.

I think this may be achievable by using 'fdupes' from the command line,
but haven't tried it yet for this and am unsure what Digikam is using
for this? Surely it must be something similar?

Regards,
-- 
Johnny



More information about the Digikam-users mailing list