FREQUENCIES on scale variables in v25

[Mike Donatello]

Prior to upgrading to v25 last week, when I ran FREQ on a varlist which included scale and non-scale variables under v24, I'd get selected numeric summaries (e.g., mean, median, skewness, etc.) for the scale variables and actual frequency tables for the nominal or ordinal variables, which is how the command is supposed to work. Since the upgrade, v25 keeps trying to give me a frequency table for every value of the scale variable, which -- naturally -- results in humongous tables and causes the procedure to die with error messages like, "Any changes made to the working file since 16-OCT-2017 15:35:11 have been lost. The time now is 11:19:34."

Does anyone else experience this behavior and/or have a means of curing it? Under v24, my syntax ran properly on a 1.2 million record file without problems, even when I named 30 or 40 variables in one FREQ statement (with a third of those being continuous measures). Under v25, I can't complete a FREQ command on the same dataset using the same syntax, even if I edit the language to include only one of the continuous measures.

Ideas?

Thanks,

MD

--

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

[Jon Peck]

That's a puzzle: AFAIK, FREQUENCIES, in all versions of Statistics, shows you counts for all variables up to the maximum number of values allowed (unless you use NOTABLE).  It pays no attention to the measurement level, although it does understand variable type.

One solution would be to use the SPSSINC SELECT VARIABLES (Utilities > Define Variable Macro) extension command to create a macro listing variables matching appropriate metadata such as the measurement level.  This command can be installed from the Extensions menu if it isn't already included.  Then you could use this defintion in appropriate commands.

On Tue, Oct 17, 2017 at 9:28 AM, Mike Donatello <[hidden email]> wrote:

Prior to upgrading to v25 last week, when I ran FREQ on a varlist which included scale and non-scale variables under v24, I'd get selected numeric summaries (e.g., mean, median, skewness, etc.) for the scale variables and actual frequency tables for the nominal or ordinal variables, which is how the command is supposed to work. Since the upgrade, v25 keeps trying to give me a frequency table for every value of the scale variable, which -- naturally -- results in humongous tables and causes the procedure to die with error messages like, "Any changes made to the working file since 16-OCT-2017 15:35:11 have been lost. The time now is 11:19:34."

Does anyone else experience this behavior and/or have a means of curing it? Under v24, my syntax ran properly on a 1.2 million record file without problems, even when I named 30 or 40 variables in one FREQ statement (with a third of those being continuous measures). Under v25, I can't complete a FREQ command on the same dataset using the same syntax, even if I edit the language to include only one of the continuous measures.

Ideas?

Thanks,

MD

--

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

[Mike Donatello]

Interesting. I knew about /FORMAT = NOTABLE, but never specified it: v24 always popped output that way, so I assumed it was the default for scale-level variables.

I'll try your solution, Jon. Thanks for the help.

MD

On Tue, Oct 17, 2017 at 11:41 AM, Jon Peck <[hidden email]> wrote:

That's a puzzle: AFAIK, FREQUENCIES, in all versions of Statistics, shows you counts for all variables up to the maximum number of values allowed (unless you use NOTABLE).  It pays no attention to the measurement level, although it does understand variable type.

One solution would be to use the SPSSINC SELECT VARIABLES (Utilities > Define Variable Macro) extension command to create a macro listing variables matching appropriate metadata such as the measurement level.  This command can be installed from the Extensions menu if it isn't already included.  Then you could use this defintion in appropriate commands.

On Tue, Oct 17, 2017 at 9:28 AM, Mike Donatello <[hidden email]> wrote:

Prior to upgrading to v25 last week, when I ran FREQ on a varlist which included scale and non-scale variables under v24, I'd get selected numeric summaries (e.g., mean, median, skewness, etc.) for the scale variables and actual frequency tables for the nominal or ordinal variables, which is how the command is supposed to work. Since the upgrade, v25 keeps trying to give me a frequency table for every value of the scale variable, which -- naturally -- results in humongous tables and causes the procedure to die with error messages like, "Any changes made to the working file since 16-OCT-2017 15:35:11 have been lost. The time now is 11:19:34."

Does anyone else experience this behavior and/or have a means of curing it? Under v24, my syntax ran properly on a 1.2 million record file without problems, even when I named 30 or 40 variables in one FREQ statement (with a third of those being continuous measures). Under v25, I can't complete a FREQ command on the same dataset using the same syntax, even if I edit the language to include only one of the continuous measures.

Ideas?

Thanks,

MD

--

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

[David Marso]

Also consider /FORMAT LIMIT (n).
That will allow the variables with fewer than n categories to present as
usual with suppresion of the scale variables.

LIMIT(n). Suppress frequency tables with more than n categories. The number
of missing and valid cases and requested statistics are displayed for
suppressed tables.



Mike Donatello wrote
> Interesting. I knew about /FORMAT = NOTABLE, but never specified it: v24
> always popped output that way, so I assumed it was the default for
> scale-level variables.
>
> I'll try your solution, Jon. Thanks for the help.
>
> MD
>
> On Tue, Oct 17, 2017 at 11:41 AM, Jon Peck &lt;

> jkpeck@

> &gt; wrote:
>
>> That's a puzzle: AFAIK, FREQUENCIES, in all versions of Statistics, shows
>> you counts for all variables up to the maximum number of values allowed
>> (unless you use NOTABLE).  It pays no attention to the measurement level,
>> although it does understand variable type.
>>
>> One solution would be to use the SPSSINC SELECT VARIABLES (Utilities >
>> Define Variable Macro) extension command to create a macro listing
>> variables matching appropriate metadata such as the measurement level.
>> This command can be installed from the Extensions menu if it isn't
>> already
>> included.  Then you could use this defintion in appropriate commands.
>>
>> On Tue, Oct 17, 2017 at 9:28 AM, Mike Donatello &lt;

> mdonatello@

> &gt;
>> wrote:
>>
>>> Prior to upgrading to v25 last week, when I ran FREQ on a varlist which
>>> included scale and non-scale variables under v24, I'd get selected
>>> numeric
>>> summaries (e.g., mean, median, skewness, etc.) for the scale variables
>>> and
>>> actual frequency tables for the nominal or ordinal variables, which is
>>> how
>>> the command is supposed to work. Since the upgrade, v25 keeps trying to
>>> give me a frequency table for every value of the scale variable, which
>>> --
>>> naturally -- results in humongous tables and causes the procedure to die
>>> with error messages like, "Any changes made to the working file since
>>> 16-OCT-2017 15:35:11 have been lost. The time now is 11:19:34."
>>>
>>> Does anyone else experience this behavior and/or have a means of curing
>>> it? Under v24, my syntax ran properly on a 1.2 million record file
>>> without
>>> problems, even when I named 30 or 40 variables in one FREQ statement
>>> (with
>>> a third of those being continuous measures). Under v25, I can't complete
>>> a
>>> FREQ command on the same dataset using the same syntax, even if I edit
>>> the
>>> language to include only one of the continuous measures.
>>>
>>> Ideas?
>>>
>>> Thanks,
>>>
>>> MD
>>>
>>> --
>>> Mike Donatello
>>> 703.582.5680 <(703)%20582-5680>
>>> ===================== To manage your subscription to SPSSX-L, send a
>>> message to

> LISTSERV@.UGA

>  (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
>>
>>
>>
>>
>> --
>> Jon K Peck
>>

> jkpeck@

>>
>>
>
>
> --
> Mike Donatello
> 703.582.5680
>
> =====================
> To manage your subscription to SPSSX-L, send a message to

> LISTSERV@.UGA

>  (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





-----
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me.
---
"Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis."
Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?"
--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

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