"flattening" groups or directories

Is it possible to get a list of all the bottom-level documents in a set of nested groups?

I have a (large) set of pdfs of articles. They are in a directory which has a separate directory for each year. For example, in Papers/Articles/1882 I have all the articles from 1882. I would like to auto-classifty all the pdfs in my database. That is, I would like to auto-classify every file at the bottom layers of the nested directories.

But it seems i hvae to get a list of files first, then select them all, then use Data/auto classify. So, how do I get a list of all the files?

I have the same question with respect to nested groups. I imported my mail archives (I use the mac mail app). This produces nested groups that reflect the nested mail folders. I want to auto-classify all the emails in the archive. For that it seems I need to get a list of all the emails at every level of the hierarchy. How do I do that???

“Get a list” – and put it where? Are you thinking of something like a smart group that would be filter out everything except the “bottom”?

I don’t think there is a simple way to do this without writing your own automation to traverse your hierarchy and find the groups at the bottom. As a kluge, you could expand all groups in 3-pane view and command-click the bottom groups to select them.

(You are aware that when you auto classify these documents they will be moved by the AI to other destinations that are difficult to predict? In other words, your hierarchical model will be deconstructed.)

that’s kinda what I suspected. I once starteds trying to do this manually as you suggest. But I have almost 30,000 messages from decades of correspondence with many, many folders. So, doing it by hand really isn’t an option.

I guess that’s a critical paradox. DTPO is best when you have LOTS of data. But with LOTS of data, you have to rely on automation, which puts you at the mercy of what is available in the software.

I suppose what I really need is a good AppleScript. But, alas, that’s beyond my current ability.

Can I use smart groups? Specifically, suppose I have a group with lots of subgroups where the “bottom” layers have pdfs. This is the format I get when I index Papers 3, for example. I can create a “flat” group with all the pdfs by using a smart group that collects all files of type pdf.

Now, can I auto-group on this smart group? Or can I create a group which contains only replicants of the items in the smart group, so I can auto-group on THAT?

Yes a smart group that finds only PDFs. This smart group is usually part of any new database.

No. Smart groups are saved searches. Auto Group does not work on smart group contents – you need to use the originals.

Both these things are easy to test for yourself, by the way.