Special characters, namely # and @, seem to be ignored in searches. This is quite unfortunate, as I use them to indicate parts of a documents title. E.g., every doc has a unique ID number in its title (assigned by an AppleScript-script), as well as the date of creation:
foobar @2008-02-01#001234.pdf
When I search for “#001234”, I’ll get all documents matching “001234”. Search for “@2008” does find everything containing “2008”. Backslash-escaping, quoting, etc. doesn’t seem to help.
Is there a way to include special characters in the search?
Am I correct that this is only true for special characters at the beginning or end of a term (that and such characters by themselves)? It seems as if something like “C@2008-02-01” would allow me to search “C@2008*”, for example.
The position of the characters doesn’t matter. Your search term is equal to “C 2008*” which means C followed by at least one separator followed by a word beginning with 2008.
It remains true that DEVONthink searches consider only alphanumeric characters. There are performance reasons for that.
Your naming scheme to add meaning to names will not work in DEVONthink searches. But of course there are many other approaches that WILL work in DEVONthink searches.
For example, use searchable alphanumeric characters such as zz instead of # to denote metadata in text strings.
And/or use tags to convey the metadata. Keyword metadata can be added to the Spotlight Comments field in the Info panel.
ok, but if i’ve a bunch of text files with tagged line/concept (@tag)?
as @ is widely used to specify a tag, seems crazy to me not search for at least for this special character in a tool born to help organize ideas…
at the moment i’m forced to have text file outside by DT (only indexed) and use others tool to mine my data.
As a note: the prefixes - commonly, # and @ - fell out of favor some time ago. They were originally implemented as a means of distinguishing Tag data from content data. This was especially true when people used Spotlight Comments for tagging. With the advent of the OpenMeta standard, then Apple’s Finder Tags, the metadata was separated from the content and the prefixes were no longer needed.