data have;
input customer_no $ card_no$;
cards;
1 00002
1 00003
1 00006
1 00008
2 00005
2 00012
3 00015
4 00020
;
run;

data wanted;
    set have;
        by customer_no;
        retain cardnum;
        if first.customer_no then cardnum=0;
        cardnum+1;
        if last.customer_no;
        keep customer_no cardnum;
run;

6# reolin

你需要加上by ×××；那个x = x+1;可以直接写成x+1；即可。本质上还是把by statement给漏了

proc sql;
        select distinct customer_no,count(*)
        from a1
        group by customer_no;
quit;

proc sort data=a1;
        by customer_no;
run;

data a2;
        set a1;
        by customer_no;
        if first.customer_no then count=1;
        else count+1;
run;

data a2;
        retain count;
        length count;
        set a1;
        by customer_no;
        if first.customer_no then count=1;
        else count=count+1;
run;

谢谢各位高手，尤其是pobel和jingju11 ！

还有一个问题，就是我运行work.have时，我再运行work.a1，但程序仍然好像是运行work.have，我在逻辑库里面删除了该文件也不行，请问对于这种情况我改怎么处理？

没看太明白， data have 那一步没有写run;?