Problem with macros in a foreach loop

[Claudio Ferrantino]

Hello,

I have a problem when I try to call a global macro within a foreach loop.

I need the number of missing values for each observations over a series of "relevant" variables. The "relevance" of these variables comes from another variable (in the example dataset, the variable "group"). For example, looking at the code I am providing below, var1, var2 and var3 are relevant for the observations falling in the group 1000.
I decided to proceed as follows:

------------------

global groups_all 1000 1001 2030 4030 // one global macro containing all the possible groups
global group_code_1000 var1 var2 var3 // a number of global macros containing the relevant variables for each group
global group_code_1001 var1 var3 var7
global group_code_2030 var5 var6 var8 var14
global group_code_4030 var10 var9 var11 var12 var14 var15

gen relevant_missings = . // the variable I need to fill

foreach group in $groups_all {
    egen miss_`group'=rowmiss($group_code_`group') // here is where the error comes from: "1000 invalid name" r(198); plus, through set trace on, it seems Stata does not recognizes I am calling a global macro within rowmiss (it seems running rowmiss(1000)
    replace relevant_missings=miss_`group' if group==`group'
    drop miss_`group'
}

------------------

Here is an example of my dataset. I have changed names due to privacy reasons.

------------------

clear
input int(year line) str4 var1 byte var2 str4(var3 var4) double var5 str4(var6 var7) byte var8 double var9 str4(var10 var11) byte var12 str4 var13 double(var14 var15) long group
2017  59 "Y   "  91 "Y   " "Y   "               100 "Y   " "Y   " 100 0 "Y   " "Y   " 0 "Y   " 10                 80 1000
2018  59 "Y   "  89 "Y   " "Y   "               100 "Y   " "Y   " 100 0 "Y   " "Y   " 0 "Y   " 10                 30 1000
2019  59 "Y   "  95 "Y   " "Y   "               100 "Y   " "Y   " 100 0 "Y   " "Y   " 0 "Y   " 10                 30 1000
2020  59 "Y   "  93 "Y   " "Y   "               100 "Y   " "Y   " 100 0 "N   " "Y   " 0 "Y   " 10  33.33000183105469 1001
2017 566 "N   "  87 "Y   " "N   " 66.66999816894531 "Y   " "Y   " 100 . "Y   " "Y   " 1 "Y   "  0  28.56999969482422 1001
2018 566 "N   "   . "Y   " "N   " 66.66999816894531 "Y   " "Y   " 100 . "N   " "Y   " 1 "Y   "  0               37.5 2030
2019 566 "N   "  82 "Y   " "N   " 66.66999816894531 "Y   " "Y   " 100 0 "Y   " "Y   " 1 "Y   "  0  28.56999969482422 2030
2017 283 "N   " 100 "N   " "N   "               100 "Y   " "Y   " 100 . "N   " "N   " 2 "Y   "  0 23.079999923706055 4030
2018 283 "N   "  94 "N   " "N   "               100 "Y   " "Y   " 100 . "Y   " "Y   " 0 "N   "  0                 30 4030
2019 283 "N   "  91 "N   " "N   "               100 "Y   " "Y   " 100 . "N   " "Y   " 0 "Y   "  0 27.270000457763672 4030
end

------------------

Please, let me know if there is a way to call a macro within a foreach loop (it may be useful for other purposes, too), and if there is a better way to do that task.
Lastly, apologies if I have not posted this question correctly (in case, please, let me know). This is my first post, and I hope everything is ok.

Thank you in advance.

[Maarten Buis]

The easiest solution is to just don't use globals, they are evil anyhow.

local groups_all 1000 1001 2030 4030 // one global macro containing all the possible groups
local group_code_1000 var1 var2 var3 // a number of global macros containing the relevant variables for each group
local group_code_1001 var1 var3 var7
local group_code_2030 var5 var6 var8 var14
local group_code_4030 var10 var9 var11 var12 var14 var15  

foreach group in $groups_all {
    egen miss_`group'=rowmiss(`group_code_`group'')
     drop miss_`group'
}

[Hemanshu Kumar]

Another solution (though I agree with #2 that globals are evil!)

change the offending line to:

egen miss_`group'=rowmiss(${group_code_`group'})

[Claudio Ferrantino]

Thank you very much! Both solutions seem to work.
Thank you also for suggesting to use local macros instead of globals. Probably, this is just a glimpse of how evil they can be.

Have a nice day