Merge if condition met on both data sets

Question

I have two datasets and only want them to merge if the meet a certain condition within both sets.

The code below is not correct.

data cohrt2_base_pre_&m2. cohrt2_base_a4lmig_&m2.;
     length Base_indicator_&m2. $20.;
     merge if pre_cb_&m3. = pre_cb_&m4. then do;
          custon.cohrt_base_pre_&m1. (in=a)
          VALIAT_B.STATPST_M1_&m2. (in=b drop=base
                                    (rename=(priceplan=Pre_priceplan_&m4.)));

    by kit_sim msisdn
    end;

Answer 1

PROC SQL可以这样做：

proc sql ;
  create table want (drop=base) as
  select a.*, b.*, b.priceplan as pre_priceplan_&m4
  from custon.cohrt_base_pre_&m1 a
       inner join
       VALIAT_B.STATPST_M1_&m2 b on a.kit_sim = b.kit_sim
                                and a.msisdn = b.msisdn
                                and pre_cb_&m3 = pre_cb_&m4 ;
quit ;

Answer 2

因此pre_cb_＆amp; m3和pre_cb＆amp; m4存在于两个数据集中，您想过滤它们吗？如果是这样，那么你可以添加一个where选项，类似于（未经测试的）：

data cohrt2_base_pre_&m2.
     cohrt2_base_a4lmig_&m2.
;
     length Base_indicator_&m2. $20.;
     merge 
          custon.cohrt_base_pre_&m1. (where=(pre_cb_&m3. = pre_cb_&m4.)
                                      in=a
                                      )
          VALIAT_B.STATPST_M1_&m2.   (where=(pre_cb_&m3. = pre_cb_&m4.) 
                                      drop=base
                                      rename=(priceplan=Pre_priceplan_&m4.)
                                      in=b
                                      );

    by kit_sim msisdn;
run;

如果两个数据集中都存在变量，则会在合并中发生碰撞，这通常是一个坏主意。另一个想法是创建每个数据集的视图，该数据集执行子集化，然后合并两个视图。这应该可以让你避免碰撞。