Date:         Thu, 5 Apr 2012 14:27:20 +0000
Reply-To:     "Keintz, H. Mark" <mkeintz@WHARTON.UPENN.EDU>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "Keintz, H. Mark" <mkeintz@WHARTON.UPENN.EDU>
Subject:      Re: Inserting blank rows
Comments: To: Randy <randistan69@HOTMAIL.COM>
In-Reply-To:  <201204051001.q354TDA9013265@waikiki.cc.uga.edu>
Content-Type: text/plain; charset="us-ascii"

Randy:

If your goal is to faciliate a regression using lagged variables within groups of records (i.e. each DATE/IDA combination is a group), then there is a better solution than putting in three completely blank records between groups (I presume this means that you expect to use a 3-record lag somewhere).

Consider just setting the value of lagged fields to missing as appropriate when beginning a new group, as in:

data want (drop=_: );
  retain _n 0;
  set have;
  by date ida;
  if first.ida then _N=1;
  else _n+1;

  L3=lag3(vara);
  L2=lag2(vara);
  L1=lag1(vara);
  array L {3} L1-L3;

  if _N<=3 then do _I= 1 to 4 - _n; L{_I}=.; end;
run;

proc reg data=want;
  by date ida;
  model ..... ;
quit;


regards,
Mark

-----Original Message-----
From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of Randy
Sent: Thursday, April 05, 2012 6:02 AM
To: SAS-L@LISTSERV.UGA.EDU
Subject: Inserting blank rows

Dear All:
   I want to insert blank rows at fixed points in the dataset.  The reason is that I want to run regressions using lagged variables and I do not want an overlap.

My Dataset is as follows:

Date            IDA       IDB    VarA
01MAR2012        1         1      4
01MAR2012        1         2      6
01MAR2012        1         3      3
01MAR2012        1         4      8
01MAR2012        1         5      2
02MAR2012        1         1      10
02MAR2012        1         2      14
02MAR2012        1         3      18
02MAR2012        1         4      4
01MAR2012        2         1      7
01MAR2012        2         2      3
01MAR2012        2         3      22
01MAR2012        2         4      14
01MAR2012        2         5      1

I want the data set to look as follows.  I want to insert three rows after each Date and IDA

Date            IDA       IDB    VarA
01MAR2012        1         1      4
01MAR2012        1         2      6
01MAR2012        1         3      3
01MAR2012        1         4      8
01MAR2012        1         5      2



02MAR2012        1         1      10
02MAR2012        1         2      14
02MAR2012        1         3      18
02MAR2012        1         4      4



01MAR2012        2         1      7
01MAR2012        2         2      3
01MAR2012        2         3      22
01MAR2012        2         4      14
01MAR2012        2         5      1

               Randy