reg过程中的问题

程序是这样的：
52   proc reg data=b;
53   model bhar= large_buy Dummy_insti ananum insti_sharept
53 ! fundsec_sharept BM1 size lowgrowth Ch_Cenratio P_Shcn1 Ch_Shcn1 Inddirpt
54   ;
55   run;


然后就出现了warning
WARNING: The variable _NAME_ or _TYPE_ exists in a data set that is not TYPE=CORR, COV, SSCP,
         etc.
那位高手给解答一下啊

吧数据传上来 ！ 我给你run一下试试

这个是我GOOGLE出来的，希望对你有帮助。

You must have produced the input dataset with PROC MEANS or PROC TRANSPOSE
or some other procedure that automatically enters new variables called
_type_ or _name_, which are expected in a dataset that is of the mentioned
types, e.g., a correlation dataset, but not expected in a "regular"
dataset PROC REG generally works with. And this leads to the additional
comment that if you happen to be running regressions on means of variables
(say produced with PROC MEANS), there is likely another, perhaps better,
way to do it without computing the summary stats first.