How do I work with information with a associated date?

When scanning e.g. delivery manifests which I might want to find either by date or by content I’m at a bit of a loss (newbie here) how to handle date best. Approaches I tried:

  • Naming it manually like: “manifest-xxx.2015-05-12” -> I might be able to find it if I know the date with 2015-05-12 or even 2015-05 or 2015 (the filename is used for search isn’t it?)
  • Having a folder structure like “Manifests” -> “2015” -> “05” (not sure how this helps finding stuff except by browsing)
  • Changing creation date
  • Some tagging scheme (which?)
  • Some “date” feature I have missed so far
  • ?

Generally I’m interested in how people deal with documents that have a “date”

I primarily use Hazel for this, with a rule that adds a date to the document name based on a date in the document. I also must add the date manually to some documents, and for that I have a Keyboard Maestro macro to add the date in the same form as my Hazel rule. I also tag many documents with a year date tag (2015) which makes it handy to set up smart groups to retrieve all document so tagged for any given year. So there are three ways I associate a date to a document, but unfortunately two of the three require a third-party app.

I was asked to post an example of the Hazel date rule(s) that I use, and linked here is a discussion of a rule that I use with my AT&T bill.

:smiley: :smiley:

Dates are super important to a large number of my documents. I store the date in the date ‘creation date’ field. Documents that are prepared outside Devon, say with Email Archiver or by one of my helpers who has prepared them by bookmarking them in Adobe, will have dates in the form YY-MM-DD at the front of the name. A simple applescript strips the date off the front of the filename and sets the’ creation date’. If I ever need to reverse the process, say to export the documents, an applescript restores the date to the filename.

I never put the date in the filename because the way DT compresses the filename field means that you lose the ability to see the start of filename. You also lose the ability to sort by filename.

The advantage of have the document date stored in the ‘creation date’ field is that you can make use of the clever date smart folders. Not many people realise, but if you make a smart folder with date search, it has a special folder icon which makes it really stand out.

Now when DT gets meta tags, date fields will get especially interesting and open up also sorts of possibilities. I would like to be able to store date ranges and various other information to Aeon timeline which I presently have to implement with clumsy kludges.


Thanks for your replies so far I learnt:

  • I should get Hazel which looks very useful on its own
  • Having the date in the filename is controversial
    [list][]is it useful besides it being visible to me? Does DEVONthink do anything with dates in the filename?
    [*]It seems to me that Creation Date is the best place to put the associated date
  • Do I understand it right that having the date there would also make tagging with the year obsolete?

Is there any Hazel integration with DEVONthink? How is the best workflow with this? e.g. for scanning with ScanSnap support, which automagically puts the scan into my Inbox (lets say I keep the generic name) do I then need to get it out again to run Hazel on it and reimport it back manually?

Doesn’t sound less work than looking at each file and editing the date manually, so I suppose there is a better way?

I wouldn’t conclude from what others wrote that this is “controversial” at all. @Frederiko has his methods and preferences, but that doesn’t mean you cannot put dates in file names – not at all.

Personally, everything I save is prefixed YYYYMMDD, if it is something that needs to be located on a list by date. If It is something that needs to be located on a list by name then I use YYYYMMDD as the suffix. I never care for (or look at) created date. Nothing right or wrong about that.

Thats not what I meant, by the original question and by “controversial”. I’m looking for ways that helps DEVONthink find my stuff. As far as I understood is that it has no knowledge of these dates put in a filename. Its only for a human to look at a directory. And with “controversial” I meant that its done different by different people.

My question was more directed towards how deal with dated documents in a way that DEVONthink “knows” about for

  1. Smart Groups
  2. Similarity and Autofiling

Having the date in the name is not a bad add on for humans but in order for DEVONthink to make use of it as a date (with range comparisons etc.) its useless. So its probably best to have both. Which a combination of Hazel and a script that puts the date from the filename into Creation-Date.

I’m not sure what you mean by ‘knowledge of these dates put in a file name’? The following may be of no use, but as you requested info as to handling files/documents by date, I’ll share below.

As you note in your original post, Devonthink is more than capable of recognising (i.e. ‘having knowledge’) dates in a file name. I work with historical documents, where the date (when a letter/memo was written, pamphlet published, etc) is a crucial element of the document’s information. Like Korm, I save the majority of my files on DTPO with the date prefixed (though I use YYYY-MM-DD, as I find that easier to read).

I have multiple folders within a database for different archival repositories (for which you might read read multiple folders for different clients), each roughly spanning the same time frame. I use Smart Groups to isolate files by date when necessary. In the attached screenshot, you can see how I isolate files relevant to the year 1787, and on another instance all files dated for the years between 1700 and 1710. Thus for example if I’m looking for a letter I know was published in that decade, but can’t remember who wrote it or in what folder of my DB it might be residing, I can scan through the Smart Group’s contents until I locate it. It’s also occasionally very useful for me to see what was published/written in a single year or range of years. Greater granularity could be added by making the parameter Name -> Matches -> 1787-05, and so on.

My file names will also include, where practical, the names of important people associated with each file, which can help when scanning lists for individuals (who might often appear in files scattered across several folders). However, I find Search better for this operation because I tend to be as brief as possible with filenames, especially since DT doesn’t allow you to ‘wrap’ filenames so that they are not obscured.

As for information contained within files that is not date specific (such as a location), or names not associated with the creation of the file (which in my case is usually the author or recipient), search is the best bet, assuming that you are working with files in which the text is recognisable.

I haven’t really used the See Also and Classify function, which might be what you’re looking for when you mention Similarity and Autofiling. I prefer not to let AI file my material for me. When adding files to my DBs (if they are not created internally), I name and then save them to a folder on my desktop which has the folder action Import to Selected Group and Delete attached. I have a shortcut to this folder in my Finder sidebar rather than the default Global Inbox.

Most of my files are named and filed manually in this way, because my needs constantly change and I quickly found that any attempt to automate the process required too much tinkering to remain up to date. As a result I’m forced to think more deeply about my files, and so tend to have a better recollection of both their content and their location. This suits me fine!