Proc sort with the nodupkey option does the following: It sorts, but only outputs the first record in each BY group. What you get depends on the initial order of the data.

Such a concept is foreign to SQL, which does not know anything about the order of rows in the table.

The closest you can get is SELECT DISTINCT (synonym: SELECT UNIQUE), which corresponds to the NODUP (not NODUPKEY) option in proc sort.

So you can do a query like create table data2 as select distinct State,County,VarA,VarB from data1 where VarA is not null order by state,country;

but that would correspond to proc sort data=Data1(keep=State County VarA VarB ) out=Data2 nodup; where VarA ne .; by State County; run;

which is not quite your original query.

Regards, Søren

On Thu, 16 Apr 2009 14:48:52 +0000, Christine Arriola <cmarriola@HOTMAIL.COM> wrote:

>Hi everyone, > > > >Can anyone tell me how to translate the following sort procedure into proc SQL? > > > >proc sort data=Data1(keep=State County VarA VarB ) out=Data2 nodupkey; > >where VarA ne .; > >by State County; > > > >I've searched the archive, but I can't seem to find an example that undups on more than one variable. > > > >Thanks, Christine > > > > > > > > >_________________________________________________________________ >Windows Live?: Keep your life in sync. >http://windowslive.com/explore?ocid=TXT_TAGLM_WL_allup_1a_explore_042009