programming a stata command to use the svy: prefix

[John Gallup]

Does anyone know of a howto or an introduction to enabling a user-written command to accept the svy: prefix? I am mainly looking for programming syntax suggestions, not substantive issues of using survey weights and calculating heteroskedastic variances.

I have read the section "Writing programs for use with svy" in the Stata .pdf manual entry [P] program properties. It says there that all programs using the svy: prefix must be eclass programs. I am creating a graphing command that first executes an estimation command which can accept the svy: prefix which I want to pass to it. Is it true that I cannot use the svy: prefix in this case (since I am not creating an eclass program - this is a graph, not an estimation, command)? Do I have to make do, for example, with an -svy- option for my command? That seems unfortunate.


Thanks,
John

[Attaullah Shah]

You can look inside the asdoc ado file and see how I have done that.
ssc install asdoc

[John Gallup]

Thanks, Attaullah. Of course! You can just use -syntax anything- and then use -gettoken- looking for ":". I was getting wrapped up in the -program ..., properties()- when using -svy:- before an estimation command, which is a different problem.