Dataset name -- make it permanent?

[Daniel Robertson]

Can the DATASET NAME property be permanently applied to a datafile, so
that the name persists after the file has been saved, closed, and reopened?

Dan R.

--
Daniel Robertson
Senior Research and Planning Associate
Institutional Research and Planning
Cornell University, Ithaca NY 14853-2801
607.255.9642 / irp.cornell.edu

[Richard Ristow]

At 01:28 PM 8/2/2007, Daniel Robertson wrote:

>Can the DATASET NAME property be permanently applied to a datafile, so
>that the name persists after the file has been saved, closed, and
>reopened?

It can't. The dataset name is a property of the Data Editor only, not
the saved file.

*However*, making DATASET NAME a file attribute wouldn't be difficult,
and wouldn't be foreign to SPSS philosophy. VARIABLE WIDTH and VARIABLE
ALIGNMENT are already attributes (attributes of variables) that are
meaningful only in the Data Editor. No reason not to add DATASET NAME.

(So I'm copying this to the SPSS suggestion line.)

Now, in SPSS 14+ you can define user-specific attributes for a file
(see DATAFILE ATTRIBUTE command). It shouldn't be too hard to write
Python code that would make the current dataset name an attribute
before saving the file, and that would assign the dataset name from
that attribute on loading the file. You'd have to write these as
subroutines, load the subroutine definitions when you start SPSS, and
replace your SAVE FILE and GET FILE statements by calls to those
subroutines, within BEGIN PROGRAM/ END PROGRAM.

I'd still like to see this added to native SPSS, though.

An alternative would be to make the filename (no path, no extension)
the dataset name when you GET a file. I've suggested this, but
apparently not formally to the suggestion line.

-Good wishes,
  Richard

[Manmit Shrimali-2]

Dear all,

Since long i am trying find out solution to following query/

I have to run custom table of spss 14 along with t-test and z-test i.e. compare mean with bonferroni and column proportion, both of them are in test statistics tab.

Problem: Currently i get the numbers and stat testing in two different table. Is there anyway we can merge the stat testing notation with numbers?

Ultimately i want each stat testing next to respective number. Any help is greatly appreciated.

Thanks.

[Peck, Jon]

I would not want to make a dataset name a system attribute, because assigning it automatically has side effects - another open file with the same dataset name would be closed without warning, and if I don't know this is happening, I could get some pretty weird behavior.

But users can certainly define an attribute for the dataset name and apply it when the file is opened if desired.  The spssaux module already has a function called
OpenDataFile, which includes the ability to assign the dataset name as a parameter.  It would be a tiny matter to extend with another argument that would be the attribute name to get a dataset name from.  If there's much interest, I'll do that when I have a chance.

Regards,
Jon Peck

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Richard Ristow
Sent: Thursday, August 02, 2007 1:11 PM
To: [hidden email]
Subject: Re: [SPSSX-L] Dataset name -- make it permanent?

At 01:28 PM 8/2/2007, Daniel Robertson wrote:

>Can the DATASET NAME property be permanently applied to a datafile, so
>that the name persists after the file has been saved, closed, and
>reopened?

It can't. The dataset name is a property of the Data Editor only, not
the saved file.

*However*, making DATASET NAME a file attribute wouldn't be difficult,
and wouldn't be foreign to SPSS philosophy. VARIABLE WIDTH and VARIABLE
ALIGNMENT are already attributes (attributes of variables) that are
meaningful only in the Data Editor. No reason not to add DATASET NAME.

(So I'm copying this to the SPSS suggestion line.)

Now, in SPSS 14+ you can define user-specific attributes for a file
(see DATAFILE ATTRIBUTE command). It shouldn't be too hard to write
Python code that would make the current dataset name an attribute
before saving the file, and that would assign the dataset name from
that attribute on loading the file. You'd have to write these as
subroutines, load the subroutine definitions when you start SPSS, and
replace your SAVE FILE and GET FILE statements by calls to those
subroutines, within BEGIN PROGRAM/ END PROGRAM.

I'd still like to see this added to native SPSS, though.

An alternative would be to make the filename (no path, no extension)
the dataset name when you GET a file. I've suggested this, but
apparently not formally to the suggestion line.

-Good wishes,
  Richard

[Peck, Jon]

There is a programmability solution for merging tables such as this.  If you can use the Python plug-in (which started with SPSS 14), take a look at the tables module on SPSS Developer Central, www.spss.com/devcentral.  For straightforward table types, you can use it pretty easily without knowing more than an ounce of Python.  More complex merges require a bit more knowledge.

HTH,
Jon Peck

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Manmit Shrimali
Sent: Thursday, August 02, 2007 1:23 PM
To: [hidden email]
Subject: [SPSSX-L] stat test merging with numbers

Dear all,

Since long i am trying find out solution to following query/

I have to run custom table of spss 14 along with t-test and z-test i.e. compare mean with bonferroni and column proportion, both of them are in test statistics tab.

Problem: Currently i get the numbers and stat testing in two different table. Is there anyway we can merge the stat testing notation with numbers?

Ultimately i want each stat testing next to respective number. Any help is greatly appreciated.

Thanks.