Eliminate not used variables

[FLuca]

Is there a quick way to eliminate from the dataset any variable that I do not use in the do file?

[Carlo Lazzaro]

FLuca:
see -drop-.

[Nick Cox]

Carlo Lazzaro is right but knowing which variables you do use in the do-file is the big deal here. That is anywhere between trivial (you just know or can quickly scan a short file) and very difficult (there can be a great deal of text in a do-file).

[Carlo Lazzaro]

Nick is right.
Note for myself: easy replies to potentially difficult questions are never fully on target!

[Nick Cox]

I guess the answer No would be unwelcome, although not the shortest answer ever on Statalist, one character long, which has been seen at least once. That answer was

C

and the question was ....

(Not answering at all is trivial here.)

[FLuca]

Unfortunately, it is a very long do file, that works on a very large dataset, where some variables have similar names.
So, I wanted to reduce the dataset to the parts of it that I am using, while avoiding to drop variables that I am indeed using.
I thought that perhaps Stata retained somewhere the information on used variables (e.g., I open a new Stata session, I open the do file, run it all at once, and Stata knows what I have used and stores somewhere this information).
Asides from avoiding mistakes, it would allow me to save lots of time.
Thanks anyway!