请教一个问题

shiyiming

[color=#400080:i5jrti5y][size=150:i5jrti5y]我有一个数据集，其中有一个变量叫做time，

我希望分析time值中间90%的数据，排除最大和最小的总共10%的数据，

我该怎么写这段程序，是在数据步里面处理，还是在proc means里面操作？谢谢！

[/size:i5jrti5y][/color:i5jrti5y]

shiyiming

To do it before analyzing the data may be one better choice. If you agree how RANK dealing with tied values, you can try proc rank to select the 90% time (each 5% for lower and upper):

[code:2px8oluw]proc rank data=Have out=rankHave(where =(timeRank not in (0 19))) groups=20;  
   var time;
   ranks timeRank;
run;[/code:2px8oluw]
JingJu