Question

我在sas sql工作。

我有一个脚本，它总是根据时间间隔生成不同数量的表。对于每一天的一个表，除了列平衡之外，列的名称是相同的。此列名称包含日期。表的名称是TableName_date，如TableName_07092017 ... TableName_31092017

MT          column      RT  AREA balance_07092017 
ACCOUNTS    balance_lcy 30      2004862772  
ACCOUNTS    balance_lcy 30  CA  121390255,8 
ACCOUNTS    balance_lcy 30  GL  323499587
ACCOUNTS    balance_lcy 30  TF  -7821721555 
C_ACCOUNTS  balance_lcy 35  CA  2,49733E+11 
C_ACCOUNTS  balance_lcy 35  NO  3748192715  


 MT         column      RT  AREA balance_08092017 
 ACCOUNTS   balance_lcy 30      -24278162321    
 ACCOUNTS   balance_lcy 30  CA  225363070.05    
 ACCOUNTS   balance_lcy 30  GL  3117815863.7
 ACCOUNTS   balance_lcy 30  TF  47914289803 
 C_ACCOUNTS balance_lcy 35  CA  37637391174 
 C_ACCOUNTS balance_lcy 35  NO  163722935.2

是否可以创建一个自动连接这些表的脚本？他们应该看起来像这样

MT          column      RT  AREA balance_07092017 balance_08092017 balance_09092017 balance_10092017....
ACCOUNTS    balance_lcy 30      2004862772        -24278162321
ACCOUNTS    balance_lcy 30  CA  121390255,8       225363070.05
ACCOUNTS    balance_lcy 30  GL  323499587         3117815863.7
ACCOUNTS    balance_lcy 30  TF  -7821721555       47914289803
C_ACCOUNTS  balance_lcy 35  CA  2,49733E+11       37637391174
C_ACCOUNTS  balance_lcy 35  NO  3748192715        163722935.2

这是创建我需要加入的表的代码

%macro sqlloop(start,end); 
   PROC SQL; 
     %DO DT_REP=&start. %TO &end.; 

        %let year=%sysfunc(year(&DT_REP.));
        %let month=%sysfunc(month(&DT_REP.));
        %let month1=%sysfunc(PUTN(&month.,z2.));
        %let day=%sysfunc(day(&DT_REP.));
        %let day1=%sysfunc(PUTN(&day.,z2.));
        %let datum= &day1.&month1.&year.;
        %put &datum.;

       CREATE TABLE DUPLICITY_BAL_&datum. as 
        select 'ACCOUNTS' as MT, 'balance_lcy' as column, rec_type, area, sum(balance_lcy) as balance_lcy, count(balance_lcy) as count
        from database.ACCOUNTS
        where version_no = 1
          and dt_rep = &DT_REP.
        group by rec_type, area

        union all
        select 'C_ACCOUNTS' as MT, 'balance_lcy' as column, rec_type, area, sum(balance_lcy) as balance_lcy, count(balance_lcy) as count
        from database.C_ACCOUNTS
        where version_no = 1
          and dt_rep = &DT_REP.
        group by rec_type, area;
     %END; 
   QUIT;
%mend; 

%sqlloop(start=21070, end=21073)

Answer 1

为什么要使用SQL呢？使用SAS代码合并多个数据集非常简单。如果要合并以TableName_开头的所有数据集，请使用:通配符以避免键入单个数据集名称。

data want ;
  merge TableName_: ;
  by MT column RT AREA ;
run;

为什么要将数据分成多个表开头？为什么不一次只生成所有日期？

%macro sqlloop(start,end); 
PROC SQL; 
  CREATE TABLE DUPLICITY_BAL as 
    select 'ACCOUNTS' as MT
         , 'balance_lcy' as column
         , dt_rep
         , rec_type
         , area
         , sum(balance_lcy) as balance_lcy
         , count(balance_lcy) as count
    from database.ACCOUNTS
    where version_no = 1
      and dt_rep between &start and &end
    group by 1,2,3,4,5

    union all
    select 'C_ACCOUNTS' as MT
         , 'balance_lcy' as column
         , dt_rep
         , rec_type
         , area
         , sum(balance_lcy) as balance_lcy
         , count(balance_lcy) as count
    from database.C_ACCOUNTS
    where version_no = 1
      and dt_rep between &start and &end
    group by 1,2,3,4,5
  ;
quit;
%mend sqlloop;

Answer 2

首先不要这样做，保持数据的长格式。保持数据长，意味着您的表结构不会不断变化。您的报告结构可能会发生变化 - 但如果数据保持较长，则从数据库角度来看更容易。宽格式也违反了“整洁数据”的原则。

附加表格并将balance_date重命名为余额，或者在所有数据集中重复相同的内容。

如果您希望使用广泛的报表使用PROC REPORT。如果您真的，真的，（并且您不应该）想要以宽格式存储它，那么您可以在使用带有ID和/或IDLABEL的PROC TRANSPOSE后转置它以创建相同的结构。

动态连接不同计数的表SAS

2 个答案: