SQL is a group processing language where as the Data Step is a Rox or observational processing language. So A join and a Merge are not the equivalent creatures. While it may be possible in some cases to mimic the Merge with a Join it isn't always the best solution. In this case the extra effort it would take to massage the input data amd the extra SQL code isn't worth the effort. We as SAS programmers have both at our disposal, the import thing is to recognize when one is better suited for the task than the other and use it.

On the other hand if this is school work your on your own.

Toby Dunn

On Tue, 8 Dec 2009 04:44:12 -0800, siyuan li <lisiyuan0753@GMAIL.COM> wrote:

>hi, I am sorry first for that I ask the same problem. >here has 2 datasets: > data a; >input code manager$; >cards; >145 max >145 xam >155 paul >165 kobe >; >run; >data b; >input code assistant$; >cards; >145 jerry >145 tracy >145 wade >155 chen >170 yi >; >run; >data c; >merge a b; >by code; >run; >the question is that I want to yield the same output to use proc sql >full join procedure, I try to use the procudure below,but it did not >yield the same output,can anyone help me? > proc sql; >select a.code,a.manager,b.assistant from a full join b >on a.code=b.code; >quit; > >thanks in advance