Thanks to Kit Baum, the -dlist- program in
a package of the same name on SSC that was
released last week has been revised.
To install, type
ssc inst dlist
or to overwrite an existing program.
A reminder of the main features follows my signature.
Martyn Sherriff kindly sent an example in which -dlist-
fell over when confronted with a variable with
name _d. This would happen with instructions
like
di _d[1]
or
local show : di <format> _d[1]
The reason is that _d is an acceptable
abbreviation of _dup() as far as -display-
is concerned, but also that the square brackets
which follow are illegal in a context in which only
parenthesised expressions are expected.
In other words, Stata sees this as
_d ... so far, so good
[1] ... not what I expected here
Similar problems can arise with other variables
whose numbers begin with underscores. Thus
suppose you have
gen _c = 42
di _c[1]
_c is an acceptable abbreviation for _continue,
which is precisely what Stata does. The next
token Stata sees is
[1]
and Stata will display
1
(and not 42!). Why it ignores the brackets
is a story for someone else at another time.
This conflict thus requires that -display-
never be shown the name of the variable,
which is ensured by evaluating it beforehand.
Thanks to Kevin Crow and Alan Riley for helpful
discussions of this detail.
There is also a new -trim- option. Note that
this works the other way round from -list-'s.
Nick
[email protected]
> Thanks to Kit Baum, a -dlist- package containing
> a command of that name has been added to SSC.
>
> Stata 8.2 is required.
>
> -dlist- arose from a question asked privately by
> Miguel Ampudia Fraile.
>
> -dlist- is a listing command. The -d- element
> is meant to evoke -describe-. The motivation
> for -dlist- is that on occasion you may wish
> to look simultaneously at variable names,
> variable values and variable labels. For example,
> variable names may be cryptic to some degree,
> especially if the dataset is new or unfamiliar to
> you or large or complicated. Ideally, a cryptic
> variable name will be made clear by a variable
> label.
>
>
> This is what -dlist- does:
>
> . dlist in 1
>
> 1.
> make Make and Model AMC Concord
> price Price 4,099
> mpg Mileage (mpg) 22
> rep78 Repair Record 1978 3
> headroom Headroom (in.) 2.5
> trunk Trunk space (cu. ft.) 11
> weight Weight (lbs.) 2,930
> length Length (in.) 186
> turn Turn Circle (ft.) 40
> displacement Displacement (cu. in.) 121
> gear_ratio Gear Ratio 3.58
> foreign Car type Domestic
>
> That might be useful for looking at just a few observations.
>
> By default, you get variable names, labels and values.
> There are a few options for tweaking output. Perhaps
> the most useful is scope for abbreviating names and
> variable labels and even suppressing them:
>
> . dlist in 1, name(0)
>
> 1.
> Make and Model AMC Concord
> Price 4,099
> Mileage (mpg) 22
> Repair Record 1978 3
> Headroom (in.) 2.5
> Trunk space (cu. ft.) 11
> Weight (lbs.) 2,930
> Length (in.) 186
> Turn Circle (ft.) 40
> Displacement (cu. in.) 121
> Gear Ratio 3.58
> Car type Domestic
*
* 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/
