请问：怎样在全样本中挑出幸存者样本

全样本的结构是：

id  year  leverage
1  1994  a
1  1995  b
1  1996  c
...  ...
2 1994  d
2 1995  e
...  ...
要达到的目的是：从全样本中挑选出幸存者样本，使得幸存者样本中的每一个id在至少连续10年内的leverage的值不等于零（也就是说选出至少连续存活了10年的id)

谢谢了哇！

data ex;
input id year leverage $;
cards;
1 1992  a
1 1993  b
1 1995  c
1 1996  d
1 1997  0
1 1998  e
2 1992  a
2 1993  b
2 1994  c
2 1996  d
2 1995  e
2 1997  f
2 1997  a
2 1998  b
2 1999  c
3 1992  a
3 1993  b
3 1994  0
3 1996  e
3 1995  e
3 1997  f
;
run;

proc sort data=ex;
by id year;
run;

data ex1(keep=id);
symbol=1;
count=0;
do until(last.id);
set ex;
by id;
if first.id then continue;
if year-lag(year)=1 and leverage ne "0" then symbol+1;
else symbol=1;
count+(symbol>=5);/*这里测试选择连续存活5年的id，改为10即可*/
if last.id and count>0 then output;
end;
;
run;

数据文件，附件。

data wanted;
   do until(last.id);
      set have;
          where leverage ne 0;
      by id year;
          y_diff=dif(year);
          if first.id then do; y_diff=.; count=0; end;
          if (first.id or y_diff=1) then count=count+1;
          else if y_diff>1 then count=1;
          if count>=10 then flag=1;
        end;
        if flag=1;
        keep id;
run;

ziyenano 发表于 2013-4-9 16:32
data ex;
input id year leverage $;
cards;

OK啦 灰常感谢~~

噢噢噢噢噢噢
谢谢的

data ex;
input id year leverage $;
cards;
1 1992  a
1 1993  b
1 1995  c
1 1996  d
1 1997  0
1 1998  e
2 1992  a
2 1993  b
2 1994  c
2 1996  d
2 1995  e
2 1997  f
2 1997  a
2 1998  b
2 1999  c
3 1992  a
3 1993  b
3 1994  0
3 1996  e
3 1995  e
3 1997  f
;
run;

proc sort data=ex;
by id year;
run;
data a;
set ex;
by id year;
retain live;
if first.id then live=0;
else live+1;
if last.id and live ge 5;

keep id live;
run;

proc sort data=ex nodupkey;
by id year;
run;