Re: st: inconsistent behavior of -ds-

[n j cox <n.j.cox@durham.ac.uk>]

Kit Baum and Scott Merryman pointed out that there is
no inconsistency here. -ds- is behaving as advertised.

But the original problem seems soluble more directly.

If you know you want to pick up the names of the third to last
variables, you can just go

ds pres1-vp1

or

unab foo : pres1-vp1

Caleb Southworth

I have been using -ds- to identify r(varlist) and then loop through 
those variables. However, I am puzzled by the behavior of -ds, has(type 
#/#)- syntax.

To illustrate what appears to be inconsistent behavior of -ds-, consider
this dataset with 9 string variables:

. desc

Contains data from uniondemoc82-3.dta
  obs:           180
 vars:            10
 size:        40,500 (99.9% of memory free)
-------------------------------------------------------------------------------
              storage  display     value
variable name   type   format      label      variable label
-------------------------------------------------------------------------------
union           str80  %80s
id              int    %8.0g
pres1           str20  %20s
treas1          str17  %17s
sect1           str20  %20s
sect2           str16  %16s
st1             str20  %20s
st2             str16  %16s
st3             str11  %11s
vp1             str19  %19s

. ds
union   id      pres1   treas1  sect1   sect2   st1     st2     st3
vp1

. ds , has(type string)
union   pres1   treas1  sect1   sect2   st1     st2     st3     vp1

. ds, has(type 2/8)

. ds , has(type 2/62)
pres1   treas1  sect1   sect2   st1     st2     st3     vp1

. ds , has(type 2/11)
st3

My goal is to have r(varlist) contain all the variables except the first
two (union and id), one of which is a string, other other numeric.

The question here is why -ds- lists some of the variables, specific
variables, or all variables.

*
*   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/