my control dataset named datanames with 5 variables as follows:

memtype	libname	memname	sic	state
DATA	WORK	AL61	61	AL
DATA	WORK	AL62	62	AL
DATA	WORK	AL71	71	AL
DATA	WORK	AL81	81	AL
DATA	WORK	AL91	91	AL
DATA	WORK	AK61	61	AK
DATA	WORK	AK62	62	AK
DATA	WORK	AK71	71	AK
DATA	WORK	AK81	81	AK
DATA	WORK	AK91	91	AK
DATA	WORK	KY61	61	KY
DATA	WORK	KY62	62	KY
DATA	WORK	KY71	71	KY
DATA	WORK	NM61	61	NM
DATA	WORK	NM62	62	NM
DATA	WORK	NM71	71	NM
DATA	WORK	NM81	81	NM
DATA	WORK	NM91	91	NM

variables memname, state and sic are all character variables;

my original sas datasets imported from excel do not include the two variables: state and sic but the names of datasets
contain the information:the first two letters are state initials and the digits followed are sic codes.

Now I want to add the two variables to the sas datasets. with state as character variable and sic as numeric. 

My macro program and call excute codes are as follows:


%macro addvar(state,sic);
	
	data &state&sic;
		length state $ 2.;
		set &state&sic;
		state="&state";
		code=input("&sic",8.)
	run;
%mend addvar;


data _null_;
	set datanames;
	call symput('state',state);
	call symput('sic',sic);
	call execute(cats('%nrstr(%addvar)(',state,sic,')'));
run;

When I run the above codes, I got the variable state with values the same as the variable memname, and code with values missing.
I don't know why?