Date:         Fri, 13 Jul 2001 12:41:25 -0500
Reply-To:     "Gregg P. Snell" <gsnell@datasavantconsulting.com>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "Gregg P. Snell" <gsnell@DATASAVANTCONSULTING.COM>
Organization: Data Savant Consulting
Subject:      Re: Questions on a very simple "Data manipulation"
Content-Type: text/plain; charset="iso-8859-1"

Hi Shane,

Here is one solution:


data test;
input firm $ X1 X2;
cards;
firm1 10 5
firm1 15 23
firm1 16 23
firm1 20 13
firm2 5 20
firm2 6 16
firm2 2 11
firm2 20 13
;
run;

proc sort data=test; by firm; run;

data new;
  set test;
  by firm;
  retain x3;
  if first.firm then do;
    x3=1;
    x4=.;
  end;
  else do;
    x3=x3+1;
    x4=mean(x1,lag(x1));
  end;
run;


HTH...

Gregg P. Snell
Data Savant Consulting
(913) 638-4640
(208) 977-1943 fax
http://www.datasavantconsulting.com

Co-chair of MWSUG 2001
http://www.mwsug.org/mw_2001/index.htm


Shane Jun <sjun@ALBANY.EDU> wrote in message
news:1b60e24d.0107130915.17ae6ce0@posting.google.com...
> Hi Everyone,
>
> A SAS beginner is struggling with a seemingly simple problem.
>
> I am dealing with a panel dataset: multiple firms with multiple
> periods.
>
> My dataset looks like this,
>
> firm     X1       X2
>
> firm1    10       5
> firm1    15       23
> firm1    16       23
> firm1    20       13
>
> firm2    5        20
> firm2    6        16
> firm2    2        11
> firm2    20       13
>
> Of course, the original dataset includes much greater number of
> obervations for each firm.
>
> I need to add two new variables in the dataset.
>
> The first one is time period. The variable will be an index variable
> starting from 1 to the last observation of each firm. (The dataset is
> arranged by the period already)
>
> I need the new variable in order to run the regression for each
> period.
>
> SO the new variable X3 will look like
>
> firm     X1       X2        X3
>
> firm1    10       5         1
> firm1    15       23        2
> firm1    16       23        3
> firm1    20       13        4
>
> firm2    5        20        1
> firm2    6        16        2
> firm2    2        11        3
> firm2    20       13        4
>
> The second variable that I need to add is the mean of X1 of two
> previous observations for each firm.
>
> This variable will start in the second obervation for each firm.
> Calling it X4, the new dataset would look like,
>
> firm     X1       X2        X3        X4
>
> firm1    10       5         1         .
> firm1    15       23        2         mean of 10 and 15
> firm1    16       23        3         mean of 15 and 16
> firm1    20       13        4         mean of 16 and 20
>
> firm2    5        20        1         .
> firm2    6        16        2         mean of 5 and 6
> firm2    2        11        3         mean of 6 and 2
> firm2    20       13        4         mean of 2 and 20
>
> WOuld anyone help me with this?
>
> Thanks in advance.
>
> Shane
>