Differences between command and do-file

[Minchul Park]

Hi, I am a beginner in Stata.
Recently, I am suffering from some errors that happen only on do-file.
For instance,

sysuse auto, clear
generate lprice = log(price)
preserve
collapse (max) max_lprice = lprice max_mpg = mpg (iqr) iqr_lprice = lprice iqr_mpg = mpg if !missing(rep78), by(rep78)
sort rep78
restore

This code works very well if I use command window line-by-line, but does not work on do-file.
(The code above is cited from "An introduction to Stata programming, CF Baum")

I have found a lot of same kind of problems.
So, My question is in a sentence that "Is there any difference between do and command window?"

Thank you for your time to read this question.

[Maarten Buis]

I ran that code without problem from the do-file editor. So what exact error message did you get?

[daniel klein]

My guess is that Minchul ran

...
preserve
collapse ...

from the do-file and expected to find the collapsed dataset. Minchul probably did not know that preserve will automatically restore the original dataset when the respective session, here: the do-file, ends.

To answer the question: execution of a do-file is considered a session that starts with do-ing (or run-ing) the do-file and end when the do-file ends; the session in which you use the command line starts when you open Stata and ends when you exit Stata.

[Minchul Park]

Dear Maarten and Daniel,

I found the fact that the preserve and restore should be ran together and same kind of problems can be solved by run the code with other lines.
In a sentence, the reason why I cannot get the result was that I ran the do-file line-by-line.
Sorry, my very rudimentary mistake and thank you for your response.