Thanks guys! The missing values are ignored by the min/max function. If we regard missing value as min value, will the idgroup way (by data_null_) still work?

On 8/13/09, Data _null_; <iebupdte@gmail.com> wrote: > I would transpose and use IDGROUP as follows. It works well since you > want the ties which causes a problem if you use SMALLEST and WHICHN > functions. > > If you have lots of data then a HASH solution will be faster, as > mentioned by Joe. > > data test; > input a b c d e; > id = put(_n_,z8.); > cards; > 1 1 5 4 4 > 2 3 8 5 . > ;;;; > run; > proc transpose; > by id; > run; > proc summary nway; > class id; > output idgroup(min(col1) out[2](col1 _name_)=min name); > run; > proc print; > run; > > > > On 8/13/09, yingtao <yingtaoliu@gmail.com> wrote: > > Hi all, > > I am trying to get the 2 minimum value and the corresponding variable names; > > > > data test; > > input a b c d e; > > cards; > > 1 1 5 4 4 > > 2 3 8 5 1 > > ; > > > > The desired data will be; > > a b c d e min1 min1_var min2 min2_var; > > 1 1 5 4 4 1 a 1 b > > 2 3 8 5 1 1 e 2 a > > ; > > > > I was trying to use sortn function. But the function mess up the variable name. > > Anybody has better idea? > > > > data need; > > set test ; > > array k[*] _numeric_; > > call sortn(of k[*]); > > min1 = k[1]; > > min1name=vname(k(1)); > > min2 = k[2]; > > run; > > > > > > Tao > > >

______________________________________________________________________ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email ______________________________________________________________________