Two Keyword Management Questions

by Josh Anon

Last week, a coworker of mine asked two great keyword questions, which have inspired this week's blog post. The first is that he's ended up with a number of identical-yet-different keywords, such as "bob," "Bob," "BOb," etc.. He was wondering if anyone knew of a good way to consolidate them to just be "Bob." Another coworker had a great suggestion:




  1. Make a smart album with all images having the keyword "bob" or "Bob" etc..

  2. Assign a new keyword "RealBob" to all of those images (make sure to then update your smart album).

  3. Use tip three from here to remove all the old keywords.

  4. Assign "Bob" to all of the images and remove "RealBob."

  5. Open up the keywords HUD and remove the unwanted keywords.



One important thing to remember is that if you have another library with images that have the unwanted keywords, when you load the library, Aperture will add those keywords back to the database. It won't reset the keywords on the images from the first library, but the unwanted words will be in your keywords list again. Be careful!



My coworker's second question was if he could make a smart album showing all images that don't have the keyword "Bob." Unfortunately, there is no "keyword is NOT" option. What I would recommend is to use IPTC keywords. When you set keywords, Aperture automatically adds them to the image's IPTC keywords field. To filter on them, add an IPTC entry to your smart album, and set its field to "keywords." Then, in the next popup, choose "does not contain."


iptcKeywords.jpg

This is where things get a bit funky. You would expect that the IPTC keywords search would be smart enough to divide up its comma-separated list of keywords into individual keywords, checking each one individually. That way, "is not" would do exact matching on keywords and "does not contain" would do substring matching. That doesn't seem to be the case (at least from my experimenting). Instead, it seems that "is not" only works if you have just one keyword, and you're pretty much forced to use "does not contain." Even weirder is that sometimes, "does not contain" doesn't seem to work! For example, in an image with the IPTC keywords field: "cook bay, hibiscus flower, mountain, palm trees, sailboat, water," when I set my smart album to all images where IPTC keywords does not contain "water," that image didn't go away. Furthermore, doing IPTC keyword searching can be slow because Aperture (to the best of my knowledge) doesn't do anything to make searching this field faster.



Yes, there are other things you can do (mostly tricks, potentially involving custom metadata fields) to make these "keyword is not" smart albums, but despite all their shortcomings, I still think IPTC keywords are the easiest option.



If you have a better method to do either of these tasks, feel free to post your thoughts in the comments!


9 Comments

Nick
2007-08-29 07:23:25
A simpler way is to bring up the keyword HUD. Use it to change each version of the keyword to the same spelling eg bob to Bob. To then merge all the Bob's together in the Aperture keyword heirarchy, simply drag each instance of the renamed keyword to the container that the proper Bob is in - Aperture will then merge the keywords. No need to delete them or make another RealBob!
Gustaf
2007-08-29 07:25:52
An alternative and perhaps easier solution to the first problem would be to move the keyword you want to change (lets say "BOb" and "Bob") one at a time to a different level in the keyword hierarchy, rename them to "bob" and then move them back to the same level as the correct keyword. Aperture will then ask if you want to merge the keywords. Answer yes and you're done.
Gustaf
2007-08-29 07:27:47
I'm too slow on the draw...
Michael Riedel
2007-08-29 08:07:47
I had a student ask how to remove a particular keyword from a group of images with different keywords, but not from other images with that keyword. I know I can delete the word from the HUD, but that deletes it from all images.


Any ideas?

Michael Riedel
2007-08-29 08:09:22
Oops I think the #3 listed may be the answer
Jenn
2007-08-29 08:16:50
Your example on the second point shows "water" as your last keyword and it does have a comma after it. However, I wonder if that is the case - I personally don't put commas at the end of a list. Perhaps you should search on "water" without the comma afterward and see if that works.
Josh Anon
2007-08-29 08:24:41
Great sugs, Nick and Gustaf! Jenn, what I typed into the filter field was "water" (no comma). I put the comma in the article for grammatical reasons. That silly punctuation inside quotations rule can sometimes make things read oddly :)
Jose Vazqez
2007-08-29 18:56:07
Thanks for the tips! I must say though, as a recent upgrader from iPhoto, I am shocked to find out that iPhoto might actually be better at this than Aperture. In iPhoto, I have a folder for pictures I have exported to flickr. I also have a smart folder with pictures with more than 3 stars and that are not in the flickr folder. That way I just select everything in the smart folder to export the new pictures to flickr. How do you do this in Aperture? The whole thing with the IPTC keywords seems a little flacky. I guess the better question would be, does flickrexport keep track of what it has previously sent or dou you have to make sure not to resend the same pictures twice?
island monkey
2007-08-30 05:36:38
Nick and Gustaf's method works like a dream. I use it because most of my keywords are nested. If all your keywords are in a flat list it takes a bit of work.

How about making the posted workflow shorter?


1. Make a smart album for all images having the duplicated keywords “bob” or “Bob” etc..
2. Open the keyword HUD and assign the correct keyword 'Bob' to all of those images.
3. Still in the keyword HUD, select and delete the unwanted keywords.



Avoids the need to create a new keyword or new keyword sub-folders.