Hi!
Currently TrID isn't very well suited to analyze text files; it can works just with that text formats that have a pretty strong characterization, like a common string at the start, or something like that.
With a different approach (statistical, for example) some more formats could be recognized, like sources code.
But, considering that:
"a";"b";"c"
"Rome";"Boston";"Paris"
or
1
could be both two valid CSV files (the latter of one record numeric field with just one numeric field), this in particular it's a pretty challenging format to recognize...
Bye!