Question

我有3个数据集A,B,C，其中包含以下变量

A:

period  region   city 

B:

period  city   Sales

C:

period   region   Sales

我的目标是使用A和B在C上进行左连接，以根据地理位置获取销售信息。我尝试按顺序执行：

/* Left joining B to A based on period and region */

proc sql;

Create table merge1 as

select l.* , r.* from 

A as l left join B as r 

on l.period = r.period and l.city=r.city;

quit;


/* Renaming Sales variable*/

data merged2;

set merge1;

rename Sales= s1;

run;



/*Doing another left join again, this time using C*/


proc sql;

create table merge3 as

select l.*,r.* from

A as l left join C as r

on l.period= r.period and l.region=r.region;

quit;

/*Replacing some of the values*/

data merge4;

set merge3;

Sales1= IFN(s1=., Sales, s1);

drop s1 Sales;

run;

我的问题是，如果有更好/更有效的方法来解决这个问题？特别是对于多个左连接，因为随着要匹配的数据集和可变数量的增加，该过程将变得非常繁琐，谢谢！

Answer 1

您可以在单个SQL过程中执行此操作。由于您有多个表，因此您必须逐个加入它们。

proc sql;
Create table merge1 as select 
    A.* , 
    B.sales as s1,
    C.sales as s2,
    coalesce(B.sales, C.sales) as Sales /*takes first non missing value*/
from A 
    left join B on (A.period = B.period and A.city = B.city)
    left join C on (A.period = C.period and A.region = C.region);
quit;

具有相同列的多个左连接

1 个答案: