Code not properly working on PASW DM book's example

[Luca Meyer-3]

I am exercising with PASW and I have found a non-properly working code on a INPUT PROGRAM to read nested file example (pag. 47-48) in the "Programming and Data Management for PASW® Statistics 18"'s book.

The code is the following:

INPUT PROGRAM. 

COMPUTE #count=0. 

- DATA LIST FIXED END=#eof 

/#yr 1 (A) #reg 3(A) #person 25 (A). 

- DO IF #eof OR #count = 1000. 

- END FILE. 

- END IF. 

- DO IF #yr='Y'. 

- REREAD. 

- DATA LIST /Year 3-6. 

- LEAVE Year. 

- ELSE IF #reg='R'. 

- REREAD. 

- DATA LIST / Region 5-15 (A). 

- LEAVE Region. 

- ELSE IF #person='P' AND UNIFORM(1000) < 500. 

- REREAD. 

- DATA LIST / SalesRep 7-17 (A) Sales 20-23. 

- END CASE. 

- COMPUTE #count=#count+1. 

- END IF. 

END INPUT PROGRAM.

BEGIN DATA 

Y 2002

  R Chicago 

      Jones        900  P 

      Gregory      400  P

  R Baton Rouge 

      Rodriguez    300  P

      Smith        333  P

      Grau         100  P

END DATA.

The claim is that the reading will stop at the end of file or when 1000 records are read, but I believe the #count is updated within another DO IF - END IF procedure and the 1.000 limit does not apply. 

I have tested my assumption setting #count=3 and removing the random selection but the software reads all 5 cases. I have tried to insert an "AND #count<=3" within every DO IF and ELSE IF statement within the second DO IF ... END IF procedure but I always have 5 cases.

How would you correct the syntax to have only 3 cases imported?

Thanks,

Luca

Luca Meyer
www.lucameyer.com
PASW Statistics v. 18.0.2 (2-apr-2010)

R version 2.9.2 (2009-08-24)
Mac OS X 10.6.3 (10D573) - kernel Darwin 10.3.0

[Rick Oliver-3]

Oops. Instead of:

COMPUTE #count=0.

try:

NUMERIC #count (f8).

I think the former was resetting #count to 0 for each case. You need to declare the scratch variable for use in the subsequent COMPUTE #count=#count+1, but you don't need to initialize it to 0, because scratch variables are automatically initialized to 0 -- and in this case explicitly setting it to 0 was causing the condition to never be met. (For testing purposes, you might want to remove the random UNIFORM condition.)

From:	Luca Meyer <[hidden email]>
To:	[hidden email]
Date:	05/11/2010 03:17 AM
Subject:	Code not properly working on PASW DM book's example
Sent by:	"SPSSX(r) Discussion" <[hidden email]>

---

I am exercising with PASW and I have found a non-properly working code on a INPUT PROGRAM to read nested file example (pag. 47-48) in the "Programming and Data Management for PASW® Statistics 18"'s book.

The code is the following:

INPUT PROGRAM.
COMPUTE #count=0.
- DATA LIST FIXED END=#eof
/#yr 1 (A) #reg 3(A) #person 25 (A).
- DO IF #eof OR #count = 1000.
- END FILE.
- END IF.
- DO IF #yr='Y'.
- REREAD.
- DATA LIST /Year 3-6.
- LEAVE Year.
- ELSE IF #reg='R'.
- REREAD.
- DATA LIST / Region 5-15 (A).
- LEAVE Region.
- ELSE IF #person='P' AND UNIFORM(1000) < 500.
- REREAD.
- DATA LIST / SalesRep 7-17 (A) Sales 20-23.
- END CASE.
- COMPUTE #count=#count+1.
- END IF.
END INPUT PROGRAM.

BEGIN DATA
Y 2002
  R Chicago
      Jones        900  P
      Gregory      400  P
  R Baton Rouge
      Rodriguez    300  P
      Smith        333  P
      Grau         100  P
END DATA.

The claim is that the reading will stop at the end of file or when 1000 records are read, but I believe the #count is updated within another DO IF - END IF procedure and the 1.000 limit does not apply.

I have tested my assumption setting #count=3 and removing the random selection but the software reads all 5 cases. I have tried to insert an "AND #count<=3" within every DO IF and ELSE IF statement within the second DO IF ... END IF procedure but I always have 5 cases.

How would you correct the syntax to have only 3 cases imported?

Thanks,
Luca

Luca Meyer
www.lucameyer.com
PASW Statistics v. 18.0.2 (2-apr-2010)
R version 2.9.2 (2009-08-24)
Mac OS X 10.6.3 (10D573) - kernel Darwin 10.3.0

[David Marso]

INPUT PROGRAM.
NUMERIC #count .
- DATA LIST FIXED END=#eof /#yr 1 (A) #reg 3(A) #person 25 (A).
- DO IF #eof OR #count = 3.
-  END FILE.
- ELSE IF #yr='Y'.
-  REREAD.
-  DATA LIST /Year 3-6.
-  LEAVE Year.
- ELSE IF #reg='R'.
-  REREAD.
-  DATA LIST / Region 5-15 (A).
-  LEAVE Region.
- ELSE IF #person='P'.
-  REREAD.
-  DATA LIST / SalesRep 7-17 (A) Sales 20-23.
-  END CASE.
-  COMPUTE #count=#count+1.
- END IF.
END INPUT PROGRAM.

On Tue, 11 May 2010 10:12:05 +0200, Luca Meyer <[hidden email]> wrote:

>I am exercising with PASW and I have found a non-properly working code on a
INPUT PROGRAM to read nested file example (pag. 47-48) in the "Programming
and Data Management for PASW� Statistics 18"'s book.

>
>The code is the following:
>
>INPUT PROGRAM.
>COMPUTE #count=0.
>- DATA LIST FIXED END=#eof
>        /#yr 1 (A) #reg 3(A) #person 25 (A).
>- DO IF #eof OR #count = 1000.
>-       END FILE.
>- END IF.
>- DO IF #yr='Y'.
>-       REREAD.
>-       DATA LIST /Year 3-6.
>-       LEAVE Year.
>- ELSE IF #reg='R'.
>-       REREAD.
>-       DATA LIST / Region 5-15 (A).
>-       LEAVE Region.
>- ELSE IF #person='P' AND UNIFORM(1000) < 500.
>-       REREAD.
>-       DATA LIST / SalesRep 7-17 (A) Sales 20-23.
>-       END CASE.
>-       COMPUTE #count=#count+1.
>- END IF.
>END INPUT PROGRAM.
>
>BEGIN DATA
>Y 2002
>  R Chicago
>      Jones        900  P
>      Gregory      400  P
>  R Baton Rouge
>      Rodriguez    300  P
>      Smith        333  P
>      Grau         100  P
>END DATA.
>
>The claim is that the reading will stop at the end of file or when 1000

records are read, but I believe the #count is updated within another DO IF -
END IF procedure and the 1.000 limit does not apply.
>
>I have tested my assumption setting #count=3 and removing the random
selection but the software reads all 5 cases. I have tried to insert an "AND
#count<=3" within every DO IF and ELSE IF statement within the second DO IF
... END IF procedure but I always have 5 cases.

>
>How would you correct the syntax to have only 3 cases imported?
>
>Thanks,
>Luca
>
>Luca Meyer
>www.lucameyer.com
>PASW Statistics v. 18.0.2 (2-apr-2010)
>R version 2.9.2 (2009-08-24)
>Mac OS X 10.6.3 (10D573) - kernel Darwin 10.3.0
>
>
>
>
>
>
>

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD