Changing specific cells with the number of the row in a loop command

[Khaled Kotb]

Hello everyone,

I'm asking how can I replace the value of a specific cell with the number of the row in a loop command
I've created a loop command to drop variables with missing values over than 20, and I want to count those variables that are dropped by generating a value called "missing_values" that supposed to have specific cells changing over the looping process.

Example:

foreach var of varlist ..... {
count if missing(`var')
if r(N)>20 drop `var'
replace missing_values=_n in (_n/_N)
}

Any help?

[Nick Cox]

I am not clear what you want here. Perhaps it is more like this:

gen missing_values = 0

foreach var of varlist ... {
      count if missing(`var')
      if r(N) > 20 {
              replace missing_values  = missing_values + missing(`var')
              drop `var'
       }
} 

What should be clear is that once have dropped a variable, it is too late to recall any information about the pattern of missing values in that variable.

Conversely, if the number of variables dropped is the focus, a variable is not needed.

local dropped = 0

foreach var of varlist ... {
      count if missing(`var')
      if r(N) > 20 {
              local dropped = `dropped' + 1 
              local gone `gone' `var' 
              drop `var'
       }
}

di "`dropped' variables dropped" 
di "`gone' dropped" 

In fact you could just accumulate the list of variables dropped and count it when done.

[Khaled Kotb]

That's amazing !!!

It worked better than I've even imagined!!

Thanks a lot, sir.

I really appreciate it.

[Khaled Kotb]

Another question if you allow me,

Why did we write local gone `gone' 'var' instead of local gone= 'var' ?

[Nick Cox]

... because I was accumulating the names of variables that had been dropped on the grounds that there might be two or more.

Your code in #4 would report only the name of the last variable dropped, which would be fine if there were only one such variable.