SPSS 14.0

[Allen]

Does anyone know if SPSS intends on fixing the error\warning message
which is generated whenever SPSS encounters a "Compute" statement in
syntax?
I guessing this is happening to others, right?

[Art Kendall]

There is example syntax below the sig block. It creates a tiny data file
and then does a compute.
Does this syntax generate an error on your machine?

Save all your current work, then open a new instance of SPSS. Make sure
that you put warnings, etc. into the output file. <edit> <options>
<viewer>. Cut-and-paste then run the syntax.
Does this demonstrate the problem you are talking about? If not please
restate your question in more detail.
Perhaps post a small data set that generates the problem.

Art
A...@DrKendall.org
Social Research Consultants

new file.
data list list/ x (f1).
begin data
1
2
3
4
5
end data
compute xsquared =x**2.
list.

[Allen]

Art, thanks for your interest. Below is an example of what I am
referencing. I work with child welfare data, so the example in this
case is the date a child enters foster care (FC) (latremyr latremmo
latremda) and the date they exit (dodfcyr dodfcmo dodfcda) and it's in
days, therefore, I divide by 30.4375 to get the result in months (The
child's lenght of time in FC). If you run this you should get the
error message I mentioned, interestingly, if you remove the sixth case
you won't the get error message (5 or less cases yields no error).

data list list/ dodfcyr dodfcmo dodfcda latremyr latremmo latremda.
begin data
2006 03 15 2002 01 01
2005 06 30 2001 04 30
2006 03 15 2002 01 01
2005 06 30 2001 04 30
2006 03 15 2002 01 01
2006 03 15 2002 01 01
end data.
compute
los=(yrmoda(dodfcyr,dodfcmo,dodfcda)-yrmoda(latremyr,latremmo,latremda))/30.4375.
exe.

[Allen]

Correction, not an error mesage rather a warning message which follows
-

>Warning # 92
>The limit of MXWARNS warnings in this data pass has been printed. Further
>warnings have been suppressed.

[Russ Sanders]

you are just missing data in some of the fields.
You can ignore the warning.

[Art Kendall]

I haven't used yrmoda in years, but have not encountered this situation
with lots of data transformation..

If I comment out the compute that uses yrmoda, I do not get the warning.

Using syntax generate by the date wizard does not get the warning.
I do get per call warning messages when I use
compute forecewarn = sqrt (dosfcyr * -1).

So it isn't all computes.

Putting an execute between the two methods of getting los does not
eliminate the warning.

My idea that it has to to with yrmoda is strengthened because mxwarns
set to 12 does not get the message.
when you have only 5 cases you have 10 calls to yrmoda so you do not
exceed 10 calls to warning messages.

A GUESS: Yrmoda was supposed to have a warning message saying that
yrmoda was obsolete. The flag to mxwarns was set on each call, but a
call was not made to a specific warning message.

If you need finer distinctions than integer months, you might try "days"
or use ctime.days and then divide by 30.4375

Art
A...@DrKendall.org
Social Research Consultants

new file.
show mxwarns.

data list list/ dodfcyr dodfcmo dodfcda latremyr latremmo latremda.
begin data
2006 03 15 2002 01 01
2005 06 30 2001 04 30
2006 03 15 2002 01 01
2005 06 30 2001 04 30
2006 03 15 2002 01 01
2006 03 15 2002 01 01
end data.

COMPUTE endofstay = DATE.DMY(dodfcda, dodfcmo, dodfcyr).
COMPUTE beginstay = DATE.DMY(latremda, latremmo, latremyr).
COMPUTE lengthofstay = DATEDIF(endofstay, beginstay, "months").
compute
los=(yrmoda(dodfcyr,dodfcmo,dodfcda)-yrmoda(latremyr,latremmo,latremda))/30.4375.
exe.

[Allen]

Thanks for your work on this one. Your not using the yrmoda function
for years reminds of what a dinosaur programmer I am. I should attempt
to come of age. By the way I guess the question is, is this "MXWARNS"
message something SPSS should address, regardless?

[Art Kendall]

I CC'd SPSS when I replied.

Art