Statalist The Stata Listserver

[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: transformations in commands

From   "Nick Cox" <>
To   <>
Subject   RE: st: transformations in commands
Date   Fri, 2 Jun 2006 19:06:01 +0100

Austin is right, it seems to me, although
I find the metaphor "knit your own" more

The deeper issue is why StataCorp don't allow
this. It would need a very substantial change
to Stata's commands, as every command that
allows a varlist would have to be assessed 
to see whether that could be changed to 
what might be called a tvarlist, i.e. 
a list of variable names and/or transformations
and variable names. 

Then there is all the error checking and diagnostics, 
e.g. what do you expect Stata to do if you ask for

list log(make) log(price) log(foreign) 

In short, much more difficult than it sounds. 


P.S. now StataCorp will make me look foolish by
announcing it's immiment. 

Austin Nichols
> You could roll your own, by writing a wrapper for each regression
> command that parses expressions, generates tempvars, and labels them
> with the relevant expression.  Don't users of R have to program their
> own stuff every time?
Feiveson, Alan H. 

> > I have been demonstrating the wonders of Stata to 
> users of R and
> > have been embarrassed by their criticism of  Stata's inability to
> > transform a variable within the syntax of a command. For 
> example reg y
> > log(x) won't work. The variable z=log(x) must be generated 
> first. I'm
> > curious as to whether other statistical software besides R 
> allows this
> > sort of thing. At any rate, I would strongly urge Stata Corp. to
> > incorporate this sort flexibility.

*   For searches and help try:

© Copyright 1996–2022 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index