ds - how to display fullnames

[Alison El Ayadi]

Dear Statalisters -

I have a large dataset and am looking to list out the names of all my string variables. I have found the command ds which has provided the ability to limit the output of the variables to string: ds, has(type string)

Unfortunately we have some longer variable names in the dataset which means that a number of them are summarized using the ~ mid variable name. For example:
cf_antibio~r
qol_entere~y
qol_startt~e

I'd like to have the full variable name output here instead, yet the fullnames option (used within d procedure) doesn't work in this command.

Does anyone have suggestions for outputting the names of all string variables from a dataset as full names?

Thanks in advance for your advice.

Best,
Alison

[Maarten Buis]

There is no guarantee that option that exist for one command also exist for another (why would it?) If you want to know what options are available after ds you just type help ds. There you see that one of the options is varwidth(#). The helpfile describes it as: "display width for variable names; default is varwidth(12)". So by default the variables names truncated to 12 characters, but you can change that with the varwidth() option.

Another way to get at this is use the fact that Stata commands often leave information behind in memory. You can see what is left behind by typing return list (or ereturn list after an estimation command like regress). ds leaves behind the list of variable names in r(varlist), so you can use that:

. clear

. input some_really_long_name

     some_re~e
  1. 1
  2. end

.
. ds, varwidth(22)
some_really_long_name

.
. ds
some_reall~e

. di r(varlist)
some_really_long_name

[Alison El Ayadi]

Thank you so much, Maarten. I appreciate your calling attention to this alternative way to obtain the output that I was after despite the option not being included within the procedure.