Date:         Fri, 14 Apr 2006 16:45:44 -0400
Reply-To:     "Howard Schreier <hs AT dc-sug DOT org>" <nospam@HOWLES.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "Howard Schreier <hs AT dc-sug DOT org>" <nospam@HOWLES.COM>
Subject:      Re: a WARNING on PROC TRANSPOSE
Content-Type: text/plain; charset=ISO-8859-1

On Fri, 14 Apr 2006 15:33:31 -0400, Rathindronath <mehedisas@YAHOO.COM>
wrote:

>I have a following Proc transpose program:
>
>
>proc transpose data = db_GRRR
>                out = db_GRR1;
>      by protocol TRT siteid subjid trtgrpn invname sexn;
>   id dstermn;
>   var sexn;
> run;
>
>
>I am getting the following warning while I am running the program:
>
>WARNING: 210 observations omitted due to missing ID values.
>
>I know the program is correct and giving me the result I expected. But
>reason it is giving the error is the Given Dataset I am working with has
>value for the ID variable DSTERMN for few rows the rest of the row don’t
>have value (for DSTERMN). In real world they are not missing they just
>should not be there for all the rows as the way it is in the below dataset
>( part of the dataset is shown below):
>
>Obs    SITEID    SUBJID    SEXN    TRTGRPN    DSTERMN    trt    protocol
> 87      02      50105      2         3          1        3     0003-11
> 88      02      50107      2         3                   3     0003-11
> 89      02      50111      1         3                   3     0003-11
> 90      02      50113      2         3                   3     0003-11
> 91      06      50122      1         3          3        3     0003-11
> 92      06      50125      1         3                   3     0003-11
> 93      06      50128      2         3                   3     0003-11
> 94      06      50132      2         3          5        3     0003-11

You seem to be saying that the missing values are expected. That's fine,
but SAS does not really care. It is mechanistic and will behave as it was
designed to behave in this situation.

>
>
>I know there is a way to avoid this kind of warning by writing a small
>statement about missing ID variable ( such as: DSTERMN ne .  / or
>something like that), But can not remember it.

That's OK. Nobody can remember all of SAS. That's why there is
documentation.

Hint: This is not specific to PROC TRANSPOSE. It is not completely
specific to missing values. You are just in need of a technique to subset
observations.

>Would you please be kind
>enough to give me an idea how can I avoid this warning with having any
>effect in the result.
>
>THANKS A LOT IN ADVANCE