variable name abbreviation in generate/replace sequence

[Charles Bondi]

There seems to be some inconsistency about how to abbreviate variable names. I would like to abbreviate DX1, DX2, etc. as DX* in logic similar to that below. Stata is letting me abbreviate in the describe command but not in the replace command. Can someone help or at least explain how I can abbreviate in the replace command? I'm using Stata 14.

. generate cardiovasc=0

. replace cardiovasc=1 if DX1=="03282"

(0 real changes made)

. drop cardiovasc

. generate cardiovasc=0

. replace cardiovasc=1 if DX*=="03282"

DX* invalid name

r(198);

. des DX*

storage display value

variable name type format label variable label

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

DX1 str5 %5s Diagnosis 1

DX2 str5 %5s Diagnosis 2

DX3 str5 %5s Diagnosis 3

DX4 str5 %5s Diagnosis 4

DX5 str5 %5s Diagnosis 5

DX6 str5 %5s Diagnosis 6

DX7 str5 %5s Diagnosis 7

DX8 str5 %5s Diagnosis 8

DX9 str5 %5s Diagnosis 9

DX10 str5 %5s Diagnosis 10

DX11 str5 %5s Diagnosis 11

DX12 str5 %5s Diagnosis 12

DX13 str5 %5s Diagnosis 13

DX14 str5 %5s Diagnosis 14

DX15 str5 %5s Diagnosis 15

[William Lisowski]

It is not inconsistent, and it is not the replace command that is the problem, it is your if clause, which would cause an error no matter where you used it.

DX* is a variable list (see the output of help varlist for further discussion).

Mathematical operators, such as "+" or "==", require two expressions, one on the left and one on the right. A variable is an expression, a variable list is not.

generate cardiovasc = 0
foreach d of varlist DX* {
    replace cardiovasc = 1 if `d'=="03282"
}

[Charles Bondi]

This was very helpful, thanks. I guess I didn't understand that * implies a variable list.