Date:         Wed, 1 Feb 2006 15:31:35 -0800
Reply-To:     Yifan Lu <ylu@ibiweb.org>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Yifan Lu <ylu@ibiweb.org>
Subject:      count problem
In-Reply-To:  <7.0.0.16.2.20060201160307.024d4810@mindspring.com>
Content-Type: text/plain; charset="us-ascii"

 Hi,

I have a very large time series data file that has 20 millions of records.
There are some variables that I am interested: PERSON, COMPANY and ACTIVITY.
Since it is a time series data, one PERSON may have several records (rows)
that indicate different time. All three variables are string variables. The
data set structure looks like this:
COMPANY         PERSON          ACTIVITY                TIME
1                       01                      n1                      2000
1                       01                      n1                      2001
1                       02                      n1                      2001
2                       03                      m1                      2000
2                       03                      m2                      2001
2                       03                      m2                      2002
2                       04                      m1                      2000
3                       05                      n1                      2000
....................................

I would like to know how many PERSON by COMPANY level and how many ACTIVITY
by PERSON level. Is there any easy way to do it?
Any suggestion is appreciated!

Yifan