Strange behaviour with FREQ command-SPSS continues to run FREQs after processing all cases

[Cleland, Patricia (EDU)]

I'm cleaning a data set off about 400,000 records. As part of the audit
trail, I'm checking the I have the right number of cases at each step
and that results of recoding and cleaning are what I expected.  One of
the ways that I'm doing this is with a FREQ command on some newly
created variables.

When I run that command, the frequency tables appear in the output
window but as pivot tables rather than as plain output.  The bar at the
bottom of the screen continues to display 'Running FREQUENCIES' after
the last case has been processed and the tables created.  At this point
the program freezes and I have to go to End Program in Task Manager to
get out. Needless to say, anything that had not been saved (such as the
output with the frequency tables that I'm interested in) is lost.

Here's the syntax that I'm using:

FREQUENCIES
  VARIABLES=ENR_DIFF CLASSES_TT
  /ORDER=  ANALYSIS .

This does not happen with other variables in the data set, including
other variables that are created in the run. It doesn't happen when I
run the syntax in V14, only in V15 and most interestingly of all, it
started in the middle of the day. One morning the program ran perfectly.
In the afternoon it started doing this! It's now been going in for
several days.

Has anyone seen anything like this or have any suggestions?

Pat

====================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

[Art Kendall]

does the behavior repeat itself after a reboot?

Art

Cleland, Patricia (EDU) wrote:

> I'm cleaning a data set off about 400,000 records. As part of the audit
> trail, I'm checking the I have the right number of cases at each step
> and that results of recoding and cleaning are what I expected.  One of
> the ways that I'm doing this is with a FREQ command on some newly
> created variables.
>
> When I run that command, the frequency tables appear in the output
> window but as pivot tables rather than as plain output.  The bar at the
> bottom of the screen continues to display 'Running FREQUENCIES' after
> the last case has been processed and the tables created.  At this point
> the program freezes and I have to go to End Program in Task Manager to
> get out. Needless to say, anything that had not been saved (such as the
> output with the frequency tables that I'm interested in) is lost.
>
> Here's the syntax that I'm using:
>
> FREQUENCIES
>   VARIABLES=ENR_DIFF CLASSES_TT
>   /ORDER=  ANALYSIS .
>
> This does not happen with other variables in the data set, including
> other variables that are created in the run. It doesn't happen when I
> run the syntax in V14, only in V15 and most interestingly of all, it
> started in the middle of the day. One morning the program ran perfectly.
> In the afternoon it started doing this! It's now been going in for
> several days.
>
> Has anyone seen anything like this or have any suggestions?
>
> Pat
>
> ===================
> 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
>
>
>

=====================
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

[Art Kendall]

I have not come across this problem, but while waiting for better info
from the list:
First see if a cache command helps.
second, see if a plain execute instead of the frequencies, has the problem

third check your syntax to see if the results of transforms are
producing huge numbers of distinct results. e.g., RND or TRUNC needed?

fourth try a Descriptives if the variable is numeric.

fifth subset the variables and see if it is peculiar to one or two new
variables.

Then
*IF* you still don't get better responses from the list, use this kludge
to narrow the problem down. Break your data into successively smaller
subsets, and process them one at a time.
This will tell you if it is due to something weird in your data.
If I recall correctly, split file would hold all of the results for the
different subgroups until they were all possible, so that is why,I am
suggesting something like this untested syntax.
compute hundredth=mod($casenum,100) +1.

split file by hundredth.
frequencies . . .
if that does not work comment out the split file and frequencies command
and put in
execute.

then
temporary.
select if hundredth le 50.
frequencies ...
temporary.
select if hundredth gt 50.
frequencies ...
* see which half of the cases has the problem.
Say the problem occurs in the first half
temporary.
select if hundredth le 25.
frequencies ...
temporary.
select if range(hundredth,26,50).
frequencies ...

Keep narrowing down until you find the problem or have excluded the
possibility that it is peculiar to a particular case.


Hope this helps.

Art Kendall
Social Research Consultants

Cleland, Patricia (EDU) wrote:

> Yes.
>
> -----Original Message-----
> From: Art Kendall [mailto:[hidden email]]
> Sent: June 8, 2008 9:59 AM
> To: Cleland, Patricia (EDU)
> Cc: [hidden email]
> Subject: Re: Strange behaviour with FREQ command-SPSS continues to run
> FREQs after processing all cases
>
> does the behavior repeat itself after a reboot?
>
> Art
>
> Cleland, Patricia (EDU) wrote:
>
>> I'm cleaning a data set off about 400,000 records. As part of the
>>
> audit
>
>> trail, I'm checking the I have the right number of cases at each step
>> and that results of recoding and cleaning are what I expected.  One of
>> the ways that I'm doing this is with a FREQ command on some newly
>> created variables.
>>
>> When I run that command, the frequency tables appear in the output
>> window but as pivot tables rather than as plain output.  The bar at
>>
> the
>
>> bottom of the screen continues to display 'Running FREQUENCIES' after
>> the last case has been processed and the tables created.  At this
>>
> point
>
>> the program freezes and I have to go to End Program in Task Manager to
>> get out. Needless to say, anything that had not been saved (such as
>>
> the
>
>> output with the frequency tables that I'm interested in) is lost.
>>
>> Here's the syntax that I'm using:
>>
>> FREQUENCIES
>>   VARIABLES=ENR_DIFF CLASSES_TT
>>   /ORDER=  ANALYSIS .
>>
>> This does not happen with other variables in the data set, including
>> other variables that are created in the run. It doesn't happen when I
>> run the syntax in V14, only in V15 and most interestingly of all, it
>> started in the middle of the day. One morning the program ran
>>
> perfectly.
>
>> In the afternoon it started doing this! It's now been going in for
>> several days.
>>
>> Has anyone seen anything like this or have any suggestions?
>>
>> Pat
>>
>> ===================
>> 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
>>
>>
>>
>>
>
>
>
>

=====================
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

[Matthew Reeder]

Patricia,

  I run on version 14 and actually encounter this quite frequently. Indeed, just this morning, I had to use Task Manager to close out of SPSS at least 5 times because of it hanging up on a given procedure. It happens basically anytime I ran a larger analysis (a sizable amount of output). Also, it seems to occur not only with with frequencies, but with several other procedures, as well (notably, correlations). There have been times when I've thought it happened on a certain variable, others when it seems a bit more random.

  You mention it happening in the middle of the day...do you have programs that run in the background at scheduled times during the day that might be contending with SPSS for system resources, or programs that perhaps run off of a network that you're on that you may not be aware of? Also, are you saving gigantic amounts of output into one output file? You may find it helpful to use multiple output files (e.g., run some procedures, save output file, close that output and open another one, run additional procedures, etc). It's a pain, but it's saved me quite a bit of frustration. You may be able to save and close output files from syntax in your version of SPSS (you can't in v.14, which is somewhat annoying - there's a macro to do it, but I've not ever grabbed it, so just as much my fault), which would make things quite a bit easier.

  Anyhow, from how you describe it, you're not alone in experiencing this. I've not seen many others post about it, though.



  Best,
  Matt

"Cleland, Patricia (EDU)" <[hidden email]> wrote:
  I'm cleaning a data set off about 400,000 records. As part of the audit
trail, I'm checking the I have the right number of cases at each step
and that results of recoding and cleaning are what I expected. One of
the ways that I'm doing this is with a FREQ command on some newly
created variables.

When I run that command, the frequency tables appear in the output
window but as pivot tables rather than as plain output. The bar at the
bottom of the screen continues to display 'Running FREQUENCIES' after
the last case has been processed and the tables created. At this point
the program freezes and I have to go to End Program in Task Manager to
get out. Needless to say, anything that had not been saved (such as the
output with the frequency tables that I'm interested in) is lost.

Here's the syntax that I'm using:

FREQUENCIES
VARIABLES=ENR_DIFF CLASSES_TT
/ORDER= ANALYSIS .

This does not happen with other variables in the data set, including
other variables that are created in the run. It doesn't happen when I
run the syntax in V14, only in V15 and most interestingly of all, it
started in the middle of the day. One morning the program ran perfectly.
In the afternoon it started doing this! It's now been going in for
several days.

Has anyone seen anything like this or have any suggestions?

Pat

====================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

=====================
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