Help, please! need to get PDF's out of database

I’ve imported all of a large collection of PDF files into DEVONthink (Pro, alpha), using the copy files to the database folder preference setting.

In DEVONthink, they are all nicely organized into a hierarchy of Year:Month:<PDF files>.

For some dumb reason, I thought they would be in the same hierarchy when I do an Open Package of the DTPro database, but I’ve found, to my dismay, that: 1) they are not in a hierarchy; and 2) there doesn’t seem to be any way to export the files, in a hierarchy or not, from DT Pro.

Does anyone know how I can get all the PDF files within my hierarchy out of the Files folder databasae, without individually picking them out by hand, preferably keeping them hierarchically organized?

One kludgy workaround I could do would be to sacrifice the hierarchy upon export (which would make me very sad) and do the following:

  • backup the database
  • select the hierarchy
  • perform repeated ungroups, so all the PDFs I want are in 1 group
  • drag the entire huge list of PDF’s to a folder on the Desktop
  • restore my database, to get the hierarchy back in the DT Pro database

The downside is that the exported files will be one huge list, instead of organized by Year:Month. Any other ideas?

Christian, any chance that dragging a Group out of DEVONthink can be changed to exporting that entire branch of the database, rather than a useless text file?

Thanks for any tips,

Joe

UPDATE: doh, I should have explored the menus more before posting. There is the rather obvious File:Export:Files and Folders option which does exactly what I want, I think, along with a whole bunch of DEVONtech_storage files.

For future reference, what are the DEVONtech_storage files useful for, after exporting?

Joe,

you got the answer yourself by trying out the menu command Export - Files/Folders. It works perfectly although there are  these DEVONtech_storage files. You can trash them (should be easy using FileBuddy or so) or just let them alive because they do not do any harm  ;D

Regards,
Maria

Come to think of it, would it be a user interface improvement to change the  behavior of dragging a Group from DEVONthink to default to exporting the Group and all its contents?

I personally do not see the use of exporting a text clipping that contains just the title of a group, but perhaps there are uses I am unaware of.

Joe

Joe, I believe the text files would be used by DT if you were to import the data back into DT (as in db rebuild by export/import). This text file would tell DT the name of the re-imported group.

This will be improved in a v1.9.x release as the current drag & drop implementation is still compatible to OS X 10.1.x Carbon applications (and drag & drop support of this ancient OS X version was really limited).

The DEVONtech_storage files contain all additional metadata, e.g. paths/URLs, comments, labels, state/visibility of checkbox, internal cross linking IDs etc. as there’s no room in the filesystem for this kind of information.

Hi Christian,

good to see you are back again, now a bit closer to Japan…

This puzzles me:

I feel like remembering that these storage files can be trashed, and I still do not see how DT can keep track of the metadata if there was the slightest change to the files or folders. Does the combination of exported files and folders only work if export and import happens without change inbetween, or does DT extract any useful piece of information from the storage files - possibly assigning wrong information when importing?

wondering…???
Maria

The metadata information from DEVONtech_storage files can still be used when importing files if the content of corresponding exported files has changed.  They’re also used when recreating group hierarchies.  But changing the names, types or locations of exported files can have side effects if they’re reimported when DEVONtech_store files still exist.  Actually, renaming a file would cause it to be treated like a new file during import, losing any of its metadata.  And DT probably ignores DEVONtech_store entries which don’t reference any files.

That’s how I think DEVONtech_store behave based examining a few.  They’re XML text files, readable with Property List Editor or any text editor/viewer.  There’s no checksum that would make changing a file’s content lose its association with its corresponding entry in a DEVONtech_store file.

sjk,

Thanks for the information. I never thought of trying to open one of these files. Will be a pleasure to have a look now. And thanks for explaining.

Maria

Scott,

thank you for the explanation of storage files - I couldn’t have done it better :slight_smile:

Thanks for the thanks, Maria and Christian.

:smiley:

btw, it was good to have a look into the files…