proc univariate problem

[jaselig]

hi guys , i have a llitle problem using proc univariate

i have a dataset with 4 variables i'm using univariate to get max, min , p99 for one of them.

later y merge a new variable in the data set. with this new variable i need to get the maximun value for p99 for each unique value from this new variable. the problem here is that i need also the correspondent MAX AND MIN for the max p99  i'm unable to get it.

proc univariate data=dat1

     noprint;

     VAR  var4;

     by var1 var2;

     output out=dat1

          pctlpts=99

          pctlpre=per_

           Min=Min

          Max=Max

               ;

RUN;

proc univariate data=dat1_ noprint;

     VAR per_99;

     by  var1 var3;

      ;

     output out=dat2

         Max=maxp99  ;

RUN;

y get a data set with var1 var3 and maxp99 but i need also get the correpondent min y max for the select maxp99

any ideas?

regards

[ballardw]

It will help to provide a few initial records of data in a data step and what the expected final output should be.

Since you are using noprint is suspect that what you are looking for may be easier with another approach.

[jaselig]

here some dummy data to work with ...

data dat1 ;

LENGTH

        var1         $4

        var2         $6

  var3         8

  var4         8

  var5         $5;

input

        var1     : $CHAR4.

        var2     : $CHAR6.

  var3     : BEST8.

  var4     : BEST8.

  var5     : $CHAR5.;

CARDS;

John store1 100 28005 city1

John store1 105 465 city2

John store1 103 12400 city3

John store2 104 15444 city4

John store2 5 45125 city5

John store2 40 45789 city6

John store2 8 9000 city7

;

RUN;

proc univariate data=dat1 noprint;

     VAR var3;

     by  var1 var2;

     output out=dat2

         Max=MAX  ;

RUN;

y get dat2

VAR1	VAR2	MAX
John	store1	105
John	store2	104

i need same data plus var4 and var 5 from data1

VAR1	VAR2	MAX	var4	var5
John	store1	105	465	city2
John	store2	104	15444	city4

thanks in advance.

[jaselig]

with this dummy data...

data dat1 ;

LENGTH

var1         $4

var2         $6

var3         8

var4         8

var5         $5;

input

var1     : $CHAR4.

var2     : $CHAR6.

var3     : BEST8.

var4     : BEST8.

var5     : $CHAR5.;

CARDS;

John store1 100 28005 city1

John store1 105 465 city2

John store1 103 12400 city3

John store2 104 15444 city4

John store2 5 45125 city5

John store2 40 45789 city6

John store2 8 9000 city7

;

RUN;

proc univariate data=dat1 noprint;

VAR var3;

by var1 var2;

output out=dat2

Max=MAX  ;

RUN;

data dat2 ;

merge dat1 dat2;

by  var1 var2;

RUN

i get dat2 as

VAR1	VAR2	MAX
John	store1	105
John	store2	104

i need same data plus VAR4 and VAR 5

VAR1	VAR2	MAX	var4	var5
John	store1	105	465	city2
John	store2	104	15444	city4

thanks in advance.

[Ksharp]

You missed a variable in MERGE.

Code: Program

data dat1 ;
LENGTH
   var1   $4
   var2   $6
var3   8
var4   8
var5   $5;
input
   var1   : $CHAR4.
   var2   : $CHAR6.
var3   : BEST8.
var4   : BEST8.
var5   : $CHAR5.;
CARDS;
John store1 100 28005 city1
John store1 105 465 city2
John store1 103 12400 city3
John store2 104 15444 city4
John store2 5 45125 city5
John store2 40 45789 city6
John store2 8 9000 city7
;
RUN;
proc univariate data=dat1 noprint ;
   VAR var3;
   by var1 var2;
   output out=dat2 Max=Max;
RUN;
proc sort data=dat1;by var1 var2 var3;run; 
data want;
merge dat1(rename=(var3=max)) dat2(in=inb);
by  var1 var2 max;
if inb;
RUN;

Xia Keshan