Manipulating Regression Table Coefficients Output with SpssClient

[Warren Chrusciel]

I recently started to work with SPSS 18.  At the same time I am learning
how to use SPSS with Python.  Specifically, I am trying to manipulate SPSS
output with SpssClient.

I wrote the following Python script.  I then went to Edit – Options  --
Scripts  -- Regressions and selected the script as a Coefficients Table
autoscript.

I want to display the last model in the coefficients table when I select
the stepwise option when running the regression command.  Up until this
point I am only able to change the output so that the first model is
displayed.

It seems to me that I am probably on the right track with lines 1 to 7 and
lines 11 to 12.  I am foundering with lines 8 to 10.

 1  import SpssClient
 2  SpssClient.StartClient()
 3  SpssScriptContext = SpssClient.GetScriptContext()
 4  SpssOutputItem = SpssScriptContext.GetOutputItem()
 5  SpssPivotTable = SpssOutputItem.GetSpecificType()
 6  SpssPivotMgr = SpssPivotTable.PivotManager()
 7  SpssDimension = SpssPivotMgr.GetRowDimension(1)
 8  NumCategories = SpssDimension.GetNumCategories()
 9  print "Number Categories", NumCategories
10  SpssDimension.SetCurrentCategory(NumCategories)
11  SpssDimension.MoveToLayer(0)
12  SpssClient.StopClient()

I have three questions:
1) What do I need to change in the above script to get it to do what I
want it to do?

2) Is using SpssClient the best way to achieve this or should I be using
spssaux or some other Python extension for SPSS?

3) What is a good way to learn how to use SpssClient?  I am spending a
fair amount of time studying the PASW® Statistics Scripting Guide, which
is available at Help  -- Programmability  -- Scripting.  However, it seems
to be taking me a long time to learn how to master new commands.  Are
there some example scripts somewhere?  Is there a guide to learning the
SpssClient rather than the Scripting Guide, which seems more like a
reference text to me?

Thank you,

Warren Chrusciel
Municipal Property Assessment Corporation
1420 Blair Place, Suite 300
Ottawa, Ontario K1J 9L8
[phone] ext. 2201

=====================
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 K Peck]

Jon Peck
SPSS, an IBM Company
[hidden email]
312-651-3435

From:	Warren Chrusciel <[hidden email]>
To:	[hidden email]
Date:	05/07/2010 08:55 AM
Subject:	[SPSSX-L] Manipulating Regression Table Coefficients Output with              SpssClient
Sent by:	"SPSSX(r) Discussion" <[hidden email]>

---

I recently started to work with SPSS 18.  At the same time I am learning
how to use SPSS with Python.  Specifically, I am trying to manipulate SPSS
output with SpssClient.

I wrote the following Python script.  I then went to Edit – Options  --
Scripts  -- Regressions and selected the script as a Coefficients Table
autoscript.

I want to display the last model in the coefficients table when I select
the stepwise option when running the regression command.  Up until this
point I am only able to change the output so that the first model is
displayed.

It seems to me that I am probably on the right track with lines 1 to 7 and
lines 11 to 12.  I am foundering with lines 8 to 10.

1  import SpssClient
2  SpssClient.StartClient()
3  SpssScriptContext = SpssClient.GetScriptContext()
4  SpssOutputItem = SpssScriptContext.GetOutputItem()
5  SpssPivotTable = SpssOutputItem.GetSpecificType()
6  SpssPivotMgr = SpssPivotTable.PivotManager()
7  SpssDimension = SpssPivotMgr.GetRowDimension(1)
8  NumCategories = SpssDimension.GetNumCategories()
9  print "Number Categories", NumCategories
10  SpssDimension.SetCurrentCategory(NumCategories)
11  SpssDimension.MoveToLayer(0)
12  SpssClient.StopClient()

I have three questions:
1) What do I need to change in the above script to get it to do what I
want it to do?
>>>My first suggestion would be to try the SPSSINC MODIFY TABLES extension command from Developer Central.  It can take plugin functions that are usually very short, because the command has done all the navigation and boiler plate stuff, so you can go to work directly on the relevant table or cell.  In this case, there is a function in the customstylefunctions.py module included with that command for hiding all but the final model in Regression.  Take a look at that and adapt if it isn't exactly what you want.  There are also little pivoting functions there.  These may help you to master the scripting apis as well.  With your script, be careful about the numbering.  It's different in the PivotManager from the other apis.


2) Is using SpssClient the best way to achieve this or should I be using
spssaux or some other Python extension for SPSS?

3) What is a good way to learn how to use SpssClient?  I am spending a
fair amount of time studying the PASW® Statistics Scripting Guide, which
is available at Help  -- Programmability  -- Scripting.  However, it seems
to be taking me a long time to learn how to master new commands.  Are
there some example scripts somewhere?  Is there a guide to learning the
SpssClient rather than the Scripting Guide, which seems more like a
reference text to me?
>>> There are something like 350 apis in SpssClient, so mastering them is a significant task.
The Scripting manual does have simple examples, but it certainly doesn't do a Tutorial for everything.

However, with the SPSSINC MODIFY TABLES and SPSSINC MODIFY OUTPUT, a great many things that you needed scripting for can now be done with regular syntax, or at least just snippets of scripting code.

Thank you,

Warren Chrusciel
Municipal Property Assessment Corporation
1420 Blair Place, Suite 300
Ottawa, Ontario K1J 9L8
[phone] ext. 2201

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

[Warren Chrusciel]

I downloaded the SPSSINC MODIFY TABLES zip file to my computer and
unzipped the contents to the suitable directory.  I then installed the
custom dialog package using Utilities>Custom Dialogs>Install Custom
Dialog.  I am able to generate the following syntax using Utilities>Modify
Table Appearance:

SPSSINC MODIFY TABLES subtype="'Coefficients'"
SELECT="Model"
DIMENSION= ROWS
LEVEL = -1  PROCESS = ALL
/STYLES  APPLYTO=BOTH
CUSTOMFUNCTION="customstylefunctions.hideNonfinalRows".

However, I don't understand all the option that are available on the
Modify Table Appearance command.  The command seems to be accessing the
correct table but as I change options it is still displaying all rows and
not just those of the final regression.  Which option(s) do I need to
change in SPSSINC MODIFY TABLES?

Thanks for all the help up to this point.

Warren Chrusciel
Municipal Property Assessment Corporation
1420 Blair Place, Suite 300
Ottawa, Ontario K1J 9L8

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