How can I use if command more efficiently to select cases among 15 consecutive variables?

[Reza Hosseini]

Hi everyone,

I am working on a large database. The database has up to 15 variables for diagnosis of the disease. They are called DX1, DX2, DX3, etc. I want to select my cases based on certain diagnosis (let's say the diagnosis code for my cases is 275). This is what I was doing:

generate MyCases=0
replace MyCases=1 if DX1=="275" | DX2="275" | DX3="275" | DX ..... | DX15="275"

I am kinda sure that there is a better and smarter way of selecting (replacing) the cases but I don't know how to do it. Also, I tried using the * after the variable as follows:

generate MyCases=0
replace MyCases=1 if DX*=="275"

but the above command gives me an error: DX ambiguous abbreviation

I appreciate if anyone knows a faster/more efficient way of doing this.

Thanks,
Reza

[Nick Cox]

Wildcards won't work in that way:

Note that this will work:

gen mycases = inlist("275", DX1, DX2, DX3)

where you can extend the list of arguments to include 10 such variables, but not all 15. So I wouldn't do it like that, even though you could split the problem into 10 and 5 variables, because it is so tedious to type out such syntax. I would do it as a loop:

gen mycases = 0

quietly forval j = 1/15 {
      replace mycases = 1 if DX`j' == "275"
}

If your values are all numeric, check out various egen functions.

[Reza Hosseini]

Wildcards won't work in that way:

Note that this will work:

gen mycases = inlist("275", DX1, DX2, DX3)

where you can extend the list of arguments to include 10 such variables, but not all 15. So I wouldn't do it like that, even though you could split the problem into 10 and 5 variables, because it is so tedious to type out such syntax. I would do it as a loop:

gen mycases = 0

quietly forval j = 1/15 {
replace mycases = 1 if DX`j' == "275"
}

Thank you so much, Nick! The loop solution makes my work much faster! Thanks again!