How many tags to an item? In a group? In a database? How deep a hierarchy?

Is there a hard limit to the number of tags that can be associated with an item? The number of tags in a group? In a database? Any limit to the hierarchical nesting of tags? If it makes a difference I’m running DEVONthink 3.6.1 on macOS 10.15 and 11.1, that’s Catalina and Big Sur for those of us who prefer words. :grinning:

Why? I’m using DEVONthink for genealogy and family history and rely heavily on tags in my workflow. There are 3 distinct aspects of traditional genealogy: research, analysis, and reporting. As each item is discovered during research, I import (or if impractical, index) it into a database and tag it with the exact data found in the item. For example, a person’s name, “Marie” is tagged exactly as “Marie”, not “Maria”, or “Mary”; a place, “Phily” would be tagged “Phily” not “Philly” or “Philadelphia” even though I may “know” one of the others is “correct”. During analysis, I can use DEVONthink’s excellent tools to classify items, find similarities between/among them, and show connections. And yes, during analysis, I add more tags, effectively mapping/associating the raw data from the research phase with the conclusions of my analysis, in the above examples, perhaps, ‘Mary’ and ‘Philadelphia’.

I would expect at some point I might see a performance issue, and that would be acceptable. In practice, so far, I’ve reached up to over 100 tags per item and over 20K tags per database without any issues whatsoever. If there is/are hard limits I could always create a subset of my data by geographic region or time period as a workaround, however it would be preferable to keep an entire project together in a single database.

Thanks in advance for any assistance!

No, there is no known hard limit and there shouldn’t be any performance issues with the numbers you’re describing.


I anticipate those numbers will grow by 1 order of magnitude, and as an upper limit 2. Any concerns in those ranges? Any feel for how deep a hierarchy?

You are anticipating 1000-10,000 tags per item and 200,000-2,000,000 tags in a database??

@CaptainEddie @BLUEFROG with anywhere near this scope and with concern of pushing boundaries, I would urge the Captain mock it up. Build a test database and use scripting to add a representative number of dummy documents, tagging each with as many tags and as many levels of hierarchy wanted. Write the code and let it rip. The see what happens.

A technique I have used on numerous occasions for dubious spreadsheet users who really needed a data base for what they were doing. Worked every time.

1 Like

A single newspaper page, especially from the late 1800s where everything seems to be fine print (guess raw paper prices were high), contain 100s of names in text. In the limiting instance an article about something like a ship sinking could list over 1000 names.

Depending upon the project, whether a family, a street, a city block, a neighborhood, a small town, sources like census records, city and telephone directories, newspapers and other publications contain many 10,000s of names and name variants. Fitting them together is like a jigsaw puzzle on steroids. It’s a task DEVONthink’s AI is well suited for rather than a relational database.

Digitization of records has radically changed the nature of genealogy, family history, and micro history.

Mock up sound like a great idea! Thanks for your advice!