我想要动态重命名大约600个变量。我使用PROC SQL创建一个包含变量名称的宏变量。现在它们看起来像:aayyy,abcjjjjj,bbcjjjjj等。我想在前2个或3个字符(取决于变量)之后添加下划线并保持其余部分相同。所以最终变量看起来像aa_yyy,abc_jjjjj,bbc_jjjjj。
有什么想法吗?
答案 0 :(得分:3)
libname anylib "E:\";
data anylib.table1;
length aayyy eeeeee abcjjjjj bbcjjjjj abcdejd 8;
run;
data work.table2;
length aayyy abcjjjjj bbcjjjjj abcdejd 8;
run;
proc sql;
create table list as
select * from (
select libname, memname, name,
case
when
compress(substr(name, 3), substr(name, 3, 1)) = ''
then catt(substr(name, 1, 2), '_', substr(name, 3))
when
compress(substr(name, 4), substr(name, 4, 1)) = ''
then catt(substr(name, 1, 3), '_', substr(name, 4))
else '' end
as newname
from dictionary.columns
where libname in ('WORK', 'ANYLIB') and length(name) >= 5
and ( substr(name, 3, 1) = substr(name, 4, 1)
or substr(name, 4, 1) = substr(name, 5, 1)
)
) where newname is not null
order by libname, memname
;
quit;
data _null_;
set list;
length stmt $200;
file "E:\renames.sas";
by libname memname;
if first.libname then do;
stmt = 'proc datasets lib=' || libname || 'nolist nodetails;';
put stmt;
end;
if first.memname then do;
stmt = 'modify ' || memname || ';';
put stmt;
stmt = 'rename';
put stmt;
end;
stmt = ' ' || name || ' = ' || newname;
put stmt;
if last.memname then do;
stmt = ';';
put stmt;
end;
if last.libname then do;
stmt = 'quit;';
put stmt;
end;
run;
%include "E:\renames.sas";
compress(substr(
...背后的想法是找到第3或第4个字符重复的名称,直到名称结尾 - compress删除此字符并生成空字符串。
然后我们在数据步骤中生成PROC DATASETS
的脚本并运行(%include)它。