As a footnote to this, note that -fs- from SSC can
collect a set of filenames for you.
What -fs- does
==============
-fs- lists the names of files in compact form. Note that files do not include
directories. With no arguments, -fs- lists names of files in the current directory
or folder. Otherwise, it follows file specifications. A file specification is a
filename or (more commonly) a set of filenames indicated by the use of wildcards
* or ?. A directory or folder prefix should be given for files outside the
current directory.
See help on -dir- or -ls- for listing the names of files with more detailed
information. Unlike either of those commands, -fs- may be given more than one file
specification. Also unlike either of those commands, -fs- shows any hidden files.
Most pertinently here, the list of filenames is returned in r(files).
Application here
================
So, in the dopiest case in which your .dta files have
a rag-bag of different names, you can go
fs *.dta
foreach f in `r(files)' {
use `f', clear
<whatever>
}
Nick
[email protected]
Neil Shephard replied to Ben Dwamena
> > I wish to empirically compare a linear and nonlinear model using
> > sequentially 12 different datasets. How may I specify a looping
> > procedure for the datasets in the manner of foreach var of varlist
> > x1-x12........instead of writing a do-file for each dataset?
>
> Assuming datasets are called data1.dta, data2.dta,
> data3.dta....data12.dta, then something like...
>
> forval x = 1/12{
> use data`x', clear
> regress [etc.]
> nl [etc.]
> [any other commands]
> }
>
> ..would allow you to do this.
>
> If your filenames are non-sequential then you would use a
> foreach loop...
>
> foreach x in abc cde efg ghi ijk klm mno{
> use data`x', clear
> regress [etc.]
> nl [etc.]
> [any other commands]
> }
*
* 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/
