-ds- was introduced first in STB-8 in 1992 and then as an official Stata
command in Stata 3.1. At the time it had one role as I recall, to give a
list of variable names _only_ and thus to act as an alternative to
-describe- with minimal (and thus compact) display. In Stata now that
role is fulfilled by -describe, simple- (-d, si-) so the wheel has
turned full circle in one sense.
-ds- was then hijacked for what is (in retrospect) a different purpose,
to identify which variables fall into a defined subset, e.g. which are
string, or which are not string. At first that was via a user-written
program -ds2- but then the main ideas were built into official Stata's
-ds- during the lifetime of Stata 8. With StataCorp's characteristic
generosity I received credit as the author, meaning that I was the last
person to do a major rewrite of the program. At the same time, I
promptly lost all control. Shortly after, StataCorp changed their mind
insofar as they withdrew it from the manual.
I too would like to see it fully documented once more, and mostly
because I think it is quite [British sense] useful and needs only brief
documentation. I think it needs a new name to match the way its purpose
has morphed from one aim to another. -varfind- was bandied around at one
time. My own guess is that StataCorp's inaction on this front arises
mostly from two rather mundane but crucial facts: (a) there are always
more important things to do and (b) people who need this will find it
sooner or later, even if there is sometimes a bittersweet pleasure in
discovering something that would have been useful earlier!
Those who rummage around will find a -ds3- and a -ds5- on SSC. -ds5- was
just a version of -ds2- for Stata 5. (The numbering mixes different
conventions, I know.) -ds3- I had forgotten about until Maarten Buis
reminded me of it recently. It does include some extra features, but in
a fit of minimalism I decided that the syntax was getting too
complicated, so they did not survive to -ds-.
Nick
[email protected]
Martin Weiss
Agreed. Many tasks simply would be impossible to carry out within a
reasonable amount of time without -ds-. I encountered it for the first
time
just by chance in a program posted on the list...
David Jacobs
Maybe I shouldn't piggy back on this thread, but I've always wondered
why such a useful command like -ds- remains undocumented.
While documentation with good examples does exist in the help file,
it still doesn't make sense to me that this command is not in the
manuals.
*
* 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/
