搜索库来追加具有特定变量的数据集

Question

我在库aa & bb中分别有2个数据集。 我的代码首先检查库中的数据集，如果它们中有特定的列变量。如果数据集具有特定的变量，那么它们被追加。 但是，当我运行我的宏它不检查库中的数据集&在test1 & test2中附加它们的值，它不会执行检查数据集的预期功能，如果它们中有变量，并返回错误符号参考& ds &未找到列表&也在& ds和&列表中显示语法错误。

可你提出任何修改...

下面

是我的代码..

%macro CHK(lib1=,lib2=,varlist=); 

%local 
list 
ds 
; 

    proc sql noprint; 
    select distinct catx(".",libname,memname) into :list separated by " " 
     from dictionary.columns 
     where libname = %upcase ("&lib1") and %upcase(name) in("&varlist") ; 
    quit; 
%put &list; 

    data test1; 
    set &list; 
    run; 

    proc sql noprint; 
    select distinct catx(".",libname,memname) into :ds separated by " " 
     from dictionary.columns 
     where libname = %upcase ("&lib2") and %upcase(name) in("&varlist") ; 
    quit; 

%put &ds; 

data test2; 
set &ds; 
run; 

%mend CHK; 

%CHK(lib1=aa,lib2=bb,varlist=%str('nam', 'DD', 'ht')); 

Answer 1

我把它看成是三个步骤：

1. 获取列表数据库中包含变量的数据集。
2. 在一个可以作为CNTLIN数据集读入PROC FORMAT的结构中创建一个变量的唯一值的数据集，以创建一个将每个值映射到顺序代码的格式。
3. 处理每个数据集，使用格式重新编码变量。

下面是一些示例代码：

%macro Recode(lib=,var=); 

%local 
    DataList 
    i 
    data_i 
; 

%*1. Get a list of datasets in the library that have the variable; 

proc sql noprint ; 
    select distinct catx(".",libname,memname) into :DataList separated by " " 
    from dictionary.columns 
    where libname = %upcase("&lib") and upcase(name) = %upcase("&var"); 
quit; 

%put >>&DataList<<; 

%*2. Read the unique values of the variable into a dataset 
    and then create a format named $recode that maps each unique value 
    to a sequential code. 
; 

proc sql; 
    create table __cntlin as 
    select "$Recode" as fmtname, &var as start, put(monotonic(),z6.) as label 
    from 
     (
     %do i=1 %to %sysfunc(countw(&datalist,%str())); 
     %let data_i=%scan(&datalist,&i,%str()); 
     %if &i>1 %then union; 
     select &var from &data_i 
     %end; 
    ) 
    ; 
quit; 

proc format cntlin=__cntlin 
      library=work 
      fmtlib 
    ; 
    select $recode; 
run; 

%*3. Use the format to recode the variable in each dataset; 

%do i=1 %to %sysfunc(countw(&datalist,%str())); 
    %let data_i=%scan(&datalist,&i,%str()); 

    data &data_i._R; 
    set &data_i; 
    &var._R=put(&var,$recode.); 
    run; 
%end; 


*Cleanup; 
proc catalog cat=work.formats; 
    delete Recode.formatc; 
quit; 
proc datasets library=work memtype=data nolist; 
    delete __cntlin; 
quit; 

%mend Recode; 

%Recode(lib=work,var=name) 

有了您的样本数据，上面的代码产生两个数据集Test1_R和Test2_R：

1454 data _null_; 
1455 set Test1_R; 
1456 put @1 Name @10 Name_R; 
1457 run; 

sam  000006 
danny 000003 
jacob 000004 
susan 000009 
sandra 000007 
vinny 000010 
alicia 000001 
NOTE: There were 7 observations read from the data set WORK.TEST1_R. 

1458 
1459 data _null_; 
1460 set Test2_R; 
1461 put @1 Name @10 Name_R; 
1462 run; 

sam  000006 
dann  000002 
jhon  000005 
susan 000009 
sandy 000008 
vinny 000010 
NOTE: There were 6 observations read from the data set WORK.TEST2_R.