SAS--拼表

libname clinic 'E:sas';
*one to one reading ;
proc sort data=sasuser.admitjune out=work.adsort;
    by id;
run;
proc sort data=sasuser.stresstest out=work.stsort;
    by id;
run;

data work.one_to_one;
    set adsort(keep= id name sex age);
    set stsort(keep=timemin maxhr);
run;

proc print data=one_to_one;
run;

*concatenate;

data concatenate(drop=resthr);
    set sasuser.stress98 sasuser.stress99; *直接拼接，行数=sum旧表行数
                                                      列数<=sum旧表列数，
                                                    变量名一样放一列（ type一样，or error）;
    if resthr < 72;
run;
proc print data=concatenate;
run;

*interleaving==cbind+sort;
proc sort data=sasuser.stress98 out=work.s98;
    by tolerance;
run;
proc sort data=sasuser.stress99 out=work.s99;
    by tolerance;
run;
data work.interleave;
    set s98 s99;     *拼表顺序：根据by先放第一个表中的单位(放完为止)，根据by再放表2中的单位，
                        若表2无，回到表1，按by 的第二个值录入;
    by tolerance;
run;
proc print data=interleave;
run;


*match-merge;
data work.merged;
    merge work.adsort work.stsort; *  旧表已sort   ，跟据by，把ID相同的放在一起,列数=旧表列数和-重复变量名次数;
    by id;      *note：merge在PDV中pointer的顺序;
run;

proc print data=work.merged;
run;

*当merge的两个数据集包含多个相同变量名时，（by只用一个），只能rename，
否则该变量内容被后面数据集的变量值重写;
* (rename=(oldname=newname)); ;

proc print data=sasuser.repertory;
run;
proc print data=sasuser.finance;
run;
proc print data=sasuser.company;
run;

data work.finrep;
   merge sasuser.repertory(rename=(date=daat2))
         sasuser.finance (rename=(name=lastname))
         sasuser.company;
   by ssn;
run;


run;
proc print data=finrep;
run;

*去掉不是共有by后变量 单位;
data work.merged2;   /*merge有22个单位，merge有20个，去掉了ad和str的第一个单位miss和2458*/
    merge work.adsort(in=ina)
          work.stsort(in=ins); 
    by id;      
    if ina and ins;  *既不是0也不是missing,if判断为true是1或者nonmissing;
run;


*drop和keep变量时，注意他们是否在data步使用;
data sasuser.merged;
  merge work.adsort
      (rename=(date=AdmitDate)
      in=inad drop=height weight)
      work.stsort
      (rename=(date=VisitDate)
      in=instr keep=id tolerance date);
   by id;
   if inad and instr;
run;
proc print data=sasuser.merged;
run;