SPSSINC SPLIT DATASET

[chris aberson]

I have a dataset that looks like this

Split   y  x
1       4  6
1       6  9
1       5  6
etc.
100   4  4
100   6  9
100   8  9

In short, I have 10,000 cases, that I've split into 100 sets. I've be able to run various analyses using the standard Split File command without issue. However, I have additional analyses (e.g., Robust Regression) that do not support the basic Split File approach.

I loaded the extension/custom dialog Split Dataset (and the other piece). Whenever I open that dialog, I get the error "All variables in the working file or in variable sets in use are inappropriate for this dialog box"

Any suggestions? As described this extension seems a great addition but I can't get it off the ground

[Jon K Peck]

Which dialog are you using that gives this message?  Is it split dataset or process files or another one?  If it is split dataset, it is likely that you have the measurement levels set wrong in your dataset.  Split requires categorical variables, so if you change the measurement level of your Split variable to nominal, the dialog should be more cooperative.

HTH,


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




From:        chris aberson <[hidden email]>
To:        [hidden email],
Date:        11/28/2012 05:12 PM
Subject:        [SPSSX-L] SPSSINC SPLIT DATASET
Sent by:        "SPSSX(r) Discussion" <[hidden email]>

---

I have a dataset that looks like this

Split   y  x
1       4  6
1       6  9
1       5  6
etc.
100   4  4
100   6  9
100   8  9

In short, I have 10,000 cases, that I've split into 100 sets. I've be able
to run various analyses using the standard Split File command without issue.
However, I have additional analyses (e.g., Robust Regression) that do not
support the basic Split File approach.

I loaded the extension/custom dialog Split Dataset (and the other piece).
Whenever I open that dialog, I get the error "All variables in the working
file or in variable sets in use are inappropriate for this dialog box"

Any suggestions? As described this extension seems a great addition but I
can't get it off the ground



--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/SPSSINC-SPLIT-DATASET-tp5716517.html
Sent from the SPSSX Discussion mailing list archive at 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

[chris aberson]

Thanks Jon -- that works.

Can I ask why is that different than the base Split File approach?

[Jon K Peck]

The built-in split files mechanism long precedes the introduction of the measurement level property for variables.  I added this requirement to the SPLIT DATASET dialog variables that define groups in order to prevent people from inadvertently choosing a continuous variable as the split criterion, which would usually result in a large number of very tiny groups.

The requirement is only applied in the dialog box.  In syntax you can do whatever you want.


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




From:        chris aberson <[hidden email]>
To:        [hidden email],
Date:        11/28/2012 10:40 PM
Subject:        Re: [SPSSX-L] SPSSINC SPLIT DATASET
Sent by:        "SPSSX(r) Discussion" <[hidden email]>

---

Thanks Jon -- that works.

Can I ask why is that different than the base Split File approach?



--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/SPSSINC-SPLIT-DATASET-tp5716517p5716519.html
Sent from the SPSSX Discussion mailing list archive at 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

[chris aberson]

Thanks Jon,

Moving to the Process Data Files Portion, I continue to receive an error 10 "Invalid Index"

There are 100 files all called something like Testsplit_1 (though 100). Everything is in the correct directory and the syntax works fine when I run it individually on files

Here is what I've got


SPSSINC PROCESS FILES INPUTDATA="C:\Users\chris\Documents\Monte Carlo\testsplit*"  
SYNTAX="C:\Users\chris\Documents\Monte Carlo\robust.sps" CONTINUEONERROR=YES
VIEWERFILE= "C:\Users\chris\Documents\Monte Carlo" CLOSEDATA=YES
MACRONAME="!JOB" LOGFILEMODE=APPEND
/MACRODEFS ITEMS.

Thanks in advance for any help you can provide