Date:         Fri, 21 Jun 2002 13:41:08 +0800
Reply-To:     YCF <s007422@CUHK.EDU.HK>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         YCF <s007422@CUHK.EDU.HK>
Organization: The Chinese University of Hong Kong
Subject:      Re: Still have problems about missing data treatment

I can't use @n as the pattern of the dataset is quite irregular.

Ian  has written a program like this:
> data w ;
>   infile cards truncover ;
>   input date :yymmdd10. s :$char7. x y $ z ;
>   if z = . and not verify ( y , " 0123456789." ) then
>   do ;
>      z = input ( y , best32. ) ;
>      y = " ." ;
>   end ;
> cards ;
>   2001-04-09  s007552  25104278                     6
>   2001-04-09  s001123  1  UC  7
>   2001-04-09  a554413  2  NA 5
> ;
>proc print;
>run;

However, it works only if there is only one column after column 4. I think a
little modification is required.

>
> |--------+----------------------->
> |        |          YCF          |
> |        |          <s007422@CUHK|
> |        |          .EDU.HK>     |
> |        |                       |
> |        |          22/06/2002   |
> |        |          02:43 AM     |
> |        |          Please       |
> |        |          respond to   |
> |        |          YCF          |
> |        |                       |
> |--------+----------------------->
>   >--------------------------------------------------------|
>   |                                                        |
>   |       To:     SAS-L@LISTSERV.UGA.EDU                   |
>   |       cc:     (bcc: Philip Crane/Policy/VWA)           |
>   |       Subject:     Still have problems about missing   |
>   |       data treatment                                   |
>   >--------------------------------------------------------|
>
>
>
>
>
> Hi all,
>     I have posted this problem before but this time I would like to
specify
> my problem more clearly. Now I have a dataset like this: (Please maximize
> the window)
> date                          id           usage1     usage2      state1
> usage3   state2
> 2001-04-09       s005513   222981144   7                               300
> 2001-04-09       s005537   12               16      NOT EXPIRED    300
NOT
> STARTED
> 2001-04-09       a004122   1233344    18
> 300   STARTED
> ......
>
> The problem is that the variables "state1" and "state2" usually contain
> missing data, and there is no "." sign but just a LONGER SPACE to denote
it.
> So how to read these data in? My expected output is like this:
>
> date                          id          usage1         usage2
> state1             usage3        state2
> 2001-04-09       s005513    222981144          7               .
> 300            .
> 2001-04-09       s005537                  12        16    NOT EXPIRED
300
> NOT STARTED
> 2001-04-09       a004122        1233344        18              .
> 300    STARTED
>
> Thanks very much for your help!
>
> YCF