aggregating on binary 0/1 variable

[Anthony Babinec]

I am running version 21 on a Windows 7 64-bit machine.

 

When I omit a break variable, I experience Matt Freeman’s bug –

X_SUM is 0.

 

When I insert a variable  equal to 1 for all rows and then use it as

a break variable, I get the expected behavior – X_SUM is 1 for all

rows and Y_SUM is 2 for all rows.

 

Tony Babinec

[hidden email]

 

[Jon K Peck]

I get the right answer without a break variable, but I'm running a development build of the next fixpack for V21, so I can't test on the gold release.


Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
[hidden email]
new phone: [phone]




From:        Anthony Babinec <[hidden email]>
To:        [hidden email],
Date:        02/06/2013 02:53 PM
Subject:        [SPSSX-L] aggregating on binary 0/1 variable
Sent by:        "SPSSX(r) Discussion" <[hidden email]>

---

I am running version 21 on a Windows 7 64-bit machine.
 
When I omit a break variable, I experience Matt Freeman’s bug –
X_SUM is 0.
 
When I insert a variable  equal to 1 for all rows and then use it as
a break variable, I get the expected behavior – X_SUM is 1 for all
rows and Y_SUM is 2 for all rows.
 
Tony Babinec
tbabinec@...
 

[David Marso]

I would suggest this be immediately reported to teksport and have them determine the scope of this issue.
Back in the day, anytime a bug arose which created *WRONG RESULTS* it was escalated!
Appears to be 21ish (are there any sub versions of 21?)
Looks like happens on range of OS? (XP_32, win7_64 ...).
While it is nice to NOT have to specifically create a BREAK variable, I can't feel comfortable distributing code which uses an implicit 'non-break' because I don't know who will be using my code downstream and incorrect results from my work are unacceptable (people pay me a lot to get it right!!!) and I certainly can't afford to be sued for a FUBAR result !!!!!
So, JoNoH... Perhaps you could steer this into the right hands?
===

Anthony Babinec wrote

I am running version 21 on a Windows 7 64-bit machine.



When I omit a break variable, I experience Matt Freeman's bug -

X_SUM is 0.



When I insert a variable  equal to 1 for all rows and then use it as

a break variable, I get the expected behavior - X_SUM is 1 for all

rows and Y_SUM is 2 for all rows.



Tony Babinec

[hidden email]