Survey Dataset failing foreach loop "no observation in subpop() subpopulation" r(461)

[Praveen Harish]

Hello all

I am using a multiyear survey dataset and tabout user written command (written by Ian watson) to export results.

I am frequently running into the problem where the loop fails and says "no observation in subpop() subpopulation" r(461). I have attached the screenshot for this thread.

As the loop was failing multiple times, I tried to use the if else command as follows

count if (CASE_SURGERY_ANY_DX ==1 & CASE_`h'_PRI ==1 & YEAR==`l' )

if `r(N)' >0 {

tabout command

}


else {

di "no subpopulation"

}

However the loop is still failing and I am not sure how to rectify it.
I was hoping to see if any of the stata experts can help me.

I have posted the code and screenshot as attachment

Thank you,
Praveen

capture log close
log using "surgery.log", replace
foreach h in  MALIG BEN   {

qui use YEAR KEY  CASE_PRO_MULTI_PRI     ///
    CASE_`h'_PRI  CASE_SURGERY_ANY_DX  using "Z:/surgery_2007.dta", clear
    
    
forvalues i=2008/2011{
append   using "Z:/surgery_`i'.dta" , keep(YEAR KEY CASE_PRO_MULTI_PRI CASE_`h'_PRI CASE_SURGERY_ANY_DX)

                     }

forvalues i=2007/2011{
        
        merge 1:1 KEY using "Z:/NIS_CORE_DESIGN_`i'.dta",keepusing(KEY  DISCWT HOSPID NIS_STRATUM YEAR) update    
        drop _merge
        
        
        merge 1:1 KEY using "Z:/NIS_CORE_ANALYSIS1_`i'.dta", keepusing(KEY  FEMALE PAY1 RACE  ELIXHAUSER_CAT) update
        drop _merge

                  }
    
*   SVYSET THE POPULATON
        
svyset [pweight= DISCWT ],strata  (NIS_STRATUM ) psu( HOSPID )

        
        * setup macros for loops
            
            levelsof YEAR , local(levels)
            local YEARlabels : value label YEAR
            local counter = 0
            local filemethod = "replace"
            local heading = ""
            
            
            * begin looping through the values of the by category
                           foreach l of local levels {
                          
                           count if (CASE_SURGERY_ANY_DX ==1 & CASE_`h'_PRI  ==1 & YEAR==`l' )

                                    if  `r(N)' >0 {
                                    
                                                if `counter' > 0 {
                                                   local filemethod = "append"
                                                   local heading = "h1(nil) h2(nil)"
                                                                 }
                                                local vlabel : label `YEARlabels' `l'

                                             tabout  FEMALE RACE PAY1 ELIXHAUSER_CAT  CASE_PRO_MULTI_PRI if YEAR == `l' & (CASE_SURGERY_ANY_DX ==1 & CASE_`h'_PRI  ==1 ) using  ///
                                                     "C:/data/`h'_count_3.xls", `filemethod' `heading' h3("YEAR: `vlabel'")   c(freq ) f(0c 1 ) svy  mi
                  
            
                                    local counter = `counter' + 1    
                                
                                    
                                                  }
                                    
                                    
                                      else      {
                                         di "no subpopulation"
                                                  }

  }
  }
        *

[Phil Schumm]

I've never used tabout, but it looks like what it is doing is calling a command with the svy, subpop() prefix repeatedly to build the table based on the varlist you specified. And in (at least) one case, the corresponding subpopulation is empty (i.e., the error message you see comes from Stata, not from tabout). I don't know if tabout has any facility for handling this condition (as I said, I've never used it), but if it doesn't, then you'll have to do this another way. My suggestion would be to construct a table of counts manually using the variables you passed to tabout, and you should see that at least one cell has a count of zero.

[Praveen Harish]

Thank you for the tip.
I got it.
Thanks again
Praveen