Removing quotation marks when changing multiple label values

[Wei Hai Deng]

Dear Statalists,

I would like to add prefixes to all values across multiple label values.
Specifically, I am customizing a dtable and am indenting categorical variables by a couple extra spaces using char(160).

Here is my example code, adding "prefix_" to each value for clarity:

clear all
sysuse bplong.dta

local categ sex when

foreach var of local categ {
    local valname : value label `var' // get name of value label for variable `var'
    quietly: label list `valname' // list variable labels
    
    local newlabel // reset local label

    forvalues i = `r(min)'/`r(max)' {
        local oldsinglelabel : label `valname' `i' // get next label value
        local newsinglelabel `i' `""prefix_`oldsinglelabel'""' // add prefix
        local newlabel `newlabel' `newsinglelabel' // local with new label values list
        }

    display "`newlabel'"
    
    label define `var'_newlabel `newlabel', replace
    label values `var' `var'_newlabel
        
    }

The problem is that each new value label includes quotation marks like so:

label list

when_newlabel:
           1 "prefix_Before"
           2 "prefix_After"
sex_newlabel:
           0 "prefix_Male"
           1 "prefix_Female"
when:
           1 Before
           2 After
sex:
           0 Male
           1 Female

I can't figure out how to remove these. There must be something wrong with my use of double quotation marks.

Alternatively, is there a better way of further indenting value labels for categorical variables within dtable or using collect?

Thank you for your time.

Best regards,
Wei

[daniel klein]

I do not fully understand the ultimate goal and I have not used dtable or collect a lot. I guess that the latter two commands are probably the place to adjust the layout.

Anyway, for changing value labels, I will point to elabel (SSC, SJ, or GitHub).

. sysuse bplong.dta
(Fictional blood-pressure data)

. elabel define * (=#) (="prefix_"+@) , modify

. label list
when:
           1 prefix_Before
           2 prefix_After
sex:
           0 prefix_Male
           1 prefix_Female
agegrp:
           1 prefix_30-45
           2 prefix_46-59
           3 prefix_60+

. 
end of do-file

[Wei Hai Deng]

Thank you Daniel! Your command perfectly fits my needs.
It for some reason returned "# not allowed in eexp" when using it on my own data with variable names in elblnamelist but switching to value label names solved it.

Always love it when a single line solves my problems.

Just to explain. I wanted to change the variable name column in my descriptive table from this:

Background
 Age
 Gender
  Male
  Female
 CRP

To this:

Background
 Age
 Gender
     Male
     Female
 CRP

And figured adding spaces to the label values would be an easy solution to move the "Male" and "Female" further right since I couldn't find that dtable supports it natively.

[daniel klein]

Thank you Daniel! Your command perfectly fits my needs.
It for some reason returned "# not allowed in eexp" when using it on my own data with variable names in elblnamelist

Well, elabel define expects a value label name (or a variable name, enclosed in parentheses). However, it does not require that the value label name refers to an already defined value label. If you feed it a name that does not match any defined value label, elabel define assumes you want to define a new value label. And, new value labels do not yet have any (integer) values to which the wildcard # could refer. Hence, you cannot use wildcards when defining new value labels. This is what the error message is pointing out.

[Wei Hai Deng]

I see, I missed the fact that variable names need to be enclosed in parantheses to refer to the linked value label name.
I couldn't figure out why the same code works on the blood-pressure data but not mine but now realize that the value label names are exactly the same as the variable names in bplong.dta, not so in my data.

Thanks again for the explanation.