I would like to set up smartgroups that act globally on all open databases, with at least one DB excluded. The criteria that are offered when specifying a smartgroup do not seem to include “database”. I can only instruct DT to either search “all databases” or just a specific one. Isn’t that odd? Am I missing something? If not, this seems like a highly desirable feature for the next-gen DT.
Here is an example where this would be really useful: I have a smartfolder that shows me all files that should still be OCR’ed. This is typically done by searching for all PDF/PS with wordcount < 10. However, in that list will be many pdfs where OCR makes no sense (handwritten stuff, pure images in pdf format etc). To exclude those, I give all such pdfs a tag “no_ocr”. I can then exclude that tag in the smartgroup. This applies to my “proper” databases. I have, however, an additional database that indexes a large amount of files that are are not carefully tagged and maintained. Basically, legacy stuff not worth importing explicitly. But I still can use DT and its excellent search features to search through that material. However, this DB will have a lot unOCR’ed files and I cannot spend the effort to label them all with “no_ocr”. The straightforward solution is of course to close that database while using this particular smartfolder showing me files that need OCR action. But nevertheless, being able to exclude specific databases from a global search would be very useful.
Surely it is the case that Global searches are done not on all databases but only those that are open? Whilst I think the facility that the OP would like is desirable, and, Christian, you can count me in, a simple work around is to close the databases you do not want to search.
Thanks, Christian. It is interesting that this never came up before. It shows how divergent opinions can be. I came across this issue already a few times in less than 2 years of DT usage.
Yes, of course, I referred only to the open databases, and indeed, I already suggested the workaround of closing the unwanted DB prior to the search in the OP. But in a multi-desktop, multi-window, multi-thread world, it is easily possible that I work concurrently on some documents in that DB that I want to exclude from search. So am I supposed to shut all that down?
I have implemented something that suits me better: In the database that I want to exclude, I put a root-level group with a short, unique name, say old1, and all other groups go into that “root group”. For search purposes, that DB now can be effectively associated with the tag old1, which can easily be excluded in a smartgroup definition.
But at the end of the day, these are workarounds. Nothing more and nothing less.
I realize that programs grow historically, and not everything is in there from the beginning, and sometimes it’s hard to add stuff later, once the architecture is reasonably complete. As general feedback (and this is not criticism, just an observation), I would naively expect that search can access every single aspect of a DB entry. That would include the name of the DB in which it resides. There are other things that would be useful. Here is my list:
global search/smartfolder with DB name as criterion
search “all” also searches tag names, so that I can find anything that has the search term “mycriterion” associated with it, be it contents or metadata, including tags (without resorting to set up an advanced search)
search includes comments in pdf files. I understand that this is a limitation in the OS X pdf toolkit, not DT (because Acrobat can do it), but one can always dream. Comments that I put into marked up pdfs are among the most “search-worthy” objects I can imagine.