Re: st: cumulative average moving through time

[David Kantor <dkantor@jhu.edu>]

At 03:59 PM 10/6/2004 -0400, Dan Egan wrote, among other things:

> by sort pid (ob):gen cave = sum(calc)/ob
>
> [...]
>
> 1) Where/When did -sum()- become an acceptable argument to
> -generate-!?!? I have only ever seen it in the context of -egen-.
> Looking at the help for -generate-, there are no arguments that are
> explicitly stated to be useable. It is only at the very bottom of the
> examples that one sees an function -uniform-  and then -sum- used with
> gen. Are the others?  I know that using many egen arguments with -gen-
> will return errors (e.g. count).
(I suppose you meant "bysort pid...".)

sum() has always been a basic function.  It is not a matter of a "argument" 
to -generate-. It's a function, and any function can appear in an 
expression (subject to type compatibility). Thus it can appear anywhere an 
expression is accepted, such as in -generate-, among others.

See -help mathfun- for details.

The egen sum() program makes use of it.  See (in the appropriate ado 
directory) _gsum.ado to see how it works.

Finally, the "functions" that -egen- accepts are just those that have 
programs written for them. (They all have "_g" as a prefix to their names. 
And you can write your own egen functions if you desire.)  Some of these 
happen to have the same name as ordinary functions such as sum(). But there 
is no explicit connection between the two.

In case you wanted to know, egen is a program that calls another program. 
If you type
 egen myvar = somefunction blah blah blah
then egen will call _gsomefunction with blah blah blah as arguments.

I hope this is useful.

David Kantor
Institute for Policy Studies
Johns Hopkins University
dkantor@jhu.edu
410-516-5404

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/