Date:         Wed, 28 Jan 2009 19:59:56 -0500
Reply-To:     Arthur Tabachneck <art297@NETSCAPE.NET>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Arthur Tabachneck <art297@NETSCAPE.NET>
Subject:      Re: Can one manipulate the value of first.whatever?
Comments: To: Joe Matise <snoopy369@GMAIL.COM>

Joe,

Much appreciated.  I thought of another way, during dinner, but also
not "quite" perfect:

data want;
  set test;
  by id;
  if lag(first.id) eq 1 then first.id=1;
  deleted=0;
  if first.id then do;
    if x eq 1 then do;
      delete;
    end;
  end;
  first_id=first.id;
run;

Art
-------
On Wed, 28 Jan 2009 18:45:42 -0600, Joe Matise <snoopy369@GMAIL.COM> wrote:

>You certainly can't manipulate it the way you suggest directly, as that
>would be inconsistent with how a data step works.
>
>The ways I could imagine doing something like that:
>
>-- Use a counter:
>data
>have;
>
> input id
>x;
>
> cards;
>
>1
>1
>
>1
>1
>
>1
>3
>
>1
>4
>
>2
>1
>
>2
>2
>
>2
>3
>
>3
>2
>
>3
>3
>
>3
>4
>
>;
>
>run;
>
>
>
>proc sort
>data=have;
>
>by
>id;
>
>run;
>
>
>
>data
>have2;
>
>set
>have;
>
>by
>id;
>
>if first.id then
>counter=0;
>
>counter+1;
>
>run;
>
>
>
>
>
>data
>want;
>
> set
>have2;
>
> by
>id;
>
> retain
>firstid;
>
> if first.id then
>firstid=1;
>
> if counter = firstid then
>do;
>
>   if x eq 1 then
>do;
>
>
>firstid=firstid+1;
>
>
>delete;
>
>
>end;
>
> end;
>
>run;
>*yields 7 records, deleting the first two records in id=1 and the first
>record in id=2.
>
>-- Use a retained tracker (same HAVE dataset):
>
>
>data
>want;
>
> set
>have;
>
> by
>id;
>
> retain
>nodelete;
>
> if first.id then
>nodelete=0;
>
> if nodelete=0 then
>do;
>
>   if x eq 1 then
>do;
>
>
>delete;
>
>
>end;
>
>   else
>nodelete=1;
>
> end;
>
>run;
>
>-- DOW loop:
>
>
>
>
>
>
>data
>want;
>
> do y=1 by 1 until (last.id or x ne
>1);
>
>         set
>have;
>
>         by
>id;
>
> end;
>
> drop
>y;
>
>run;
>
>though I'm not 100% sure why this works, I was expecting it to not work at
>all, but it at least works for the example (and also for the example where
>x=1 further down from first.id, it keeps that record).  But it does seem
to
>work.
>
>-- Perhaps use LAG, though I can't come up with a functional example.
>
>-Joe
>
>
>
>On Wed, Jan 28, 2009 at 6:06 PM, Arthur Tabachneck
<art297@netscape.net>wrote:
>
>> Akshaya asked a question earlier today that made me think of a solution
>> that required a manipulative first.variable feature.  However, I don't
>> know how one could accomplish that.  For example, given the data file:
>>
>> data have;
>>  input id x;
>>  cards;
>> 1 1
>> 1 2
>> 1 3
>> 1 4
>> 2 1
>> 2 2
>> 2 3
>> 3 2
>> 3 3
>> 3 4
>> ;
>>
>> how can one, or can one, reset first.id if, because of a delete, it
isn't
>> the first record anymore?  For example:
>>
>> data want;
>>  set test;
>>  by id;
>>  if first.id then do;
>>    if x eq 1 then do;
>>      delete;
>>      /* reset first.id here*/;
>>    end;
>>  end;
>>  first_id=first.id;
>> run;
>>
>> would (if the missing command were added) end up looking like:
>>
>> data desired_want;
>>  input id x first_id;
>>  cards;
>> 1 2 1
>> 1 3 0
>> 1 4 0
>> 2 2 1
>> 2 3 0
>> 3 2 1
>> 3 3 0
>> 3 4 0
>> ;
>>
>> Just wondering,
>> Art
>>