Date: Fri, 16 Dec 2011 21:49:05 +0100
Reply-To: Jara Kampmann <jarakampmann@gmx.de>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: Jara Kampmann <jarakampmann@gmx.de>
Subject: Re: Overwriting a dataset
Content-Type: multipart/alternative;
I totally agree with all the warnings and the precautions one should take.
Nevertheless, I am wondering whether there is a possibility to save changes into the active data set via syntax. My annoyance is: if I create an outfile with aggregate, get the file and perform some changes, I always have to create a new file (with new data set name) to be able to save the changes. Which causes additional work of rewriting when having syntax to be run on several data sets as well as the need for more storing capacity.
So does one of you know of any syntax snippet that would do the trick (if necessary)?
Jara
----- Ursprüngliche Nachricht -----
Von: Jon K Peck
Gesendet: 16.12.11 21:13 Uhr
An: SPSSX-L@LISTSERV.UGA.EDU
Betreff: Re: Overwriting a dataset
Apropos of which, I would remind people that Statistics has a feature to set file permissions.
Using SAVE, you can include /PERMISSIONS=READONLY
or you can use the PERMISSIONS command:
PERMISSIONS FILE=filespec /PERMISSIONS READONLY.
If you are using C & DS (Collaboration and Deployment Services), if you save to the repository, previous generations of the file are preserved, so you can go back to earlier versions. That works for syntax and other file types, too. Preserving previous versions of your syntax file can be as important as preserving your data.
Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
peck@us.ibm.com
new phone: 720-342-5621
From: Art Kendall <Art@DrKendall.org>
To: SPSSX-L@listserv.uga.edu
Date: 12/16/2011 01:01 PM
Subject: Re: [SPSSX-L] Overwriting a dataset
Sent by: "SPSSX(r) Discussion" <SPSSX-L@listserv.uga.edu>
-----------------------------------------------------------------
I have seen things like this happen dozens of times over the years. Losing 3 hours is a pain. The worse situation I have seen was when someone wiped her data out with an "IF (something impossible);" in SAS. (SAS's version of SELECT IF). This was an hour after she shredded the paper copy of her questionnaires to protect privacy. Two years work gone.
In 1971 after a couple of weeks experience, I realized that computers do what they are told. Which is often not what you want them to do.
Art Kendall
Social Research Consultants
On 12/16/2011 2:06 PM, David Marso wrote:
One should *NEVER* overwrite data unless one can recreate it from original
sources or some well defined documented processing milestone. Sometimes
accidents happen and it can be *VERY EXPENSIVE* to recreate a crucial data
set. Once upon a time in a land far far away I was working on a project
with 4 other consultants. We had one shared dataset which was built from
several other files after sorts/merges/complex flag calculations etc... All
in all it took about 3 hours to build the master file from source data using
the complex data preparation syntax. It took about an HOUR to sort the
damned thing (500K cases, 400+ variables .. back in mid 90's when EVERYTHING
was *SLOOOWWW*).
One afternoon I grabbed the data, ran a bit of code for the part I was
working on and was QUITE distressed to find there were only 10000 cases in
the file. Turns out that one of the team members had closed out of SPSS and
clicked the F'ing button asking Save changes before closing....
Consequently 4 people were sitting on their a$$es for 3 hours while I
rebuilt the file from source files.
After that, EVERYTHING got locked down (set ALL file permissions to be READ
only).
Art Kendall wrote
Dangerous! If you overwrite a dataset will you
still be able to go back and do the process over in the way you
want versus the way you said to do it?
Art Kendall
Social Research Consultants
On 12/16/2011 11:38 AM, paulm wrote:
Hi, I am somewhat new to SPSS and I am trying to resolve an issue
that I
have when I run my syntax on the SPSS Server vs. running the syntax
locally.
Basically I am retrieving a dataset, doing some manipulations and then
saving the dataset (overwriting it). Here is the syntax I am using:
DEFINE !newdate (!positional = !TOKENS(1) / !positional = !TOKENS(1))
* RETRIEVE THE CURRENT DATASET.
GET FILE=!QUOTE(!CONCAT('e:\se\',!1,'\',!2,'\Data\POST\post_data.sav')).
DATASET NAME DataSet1 WINDOW=FRONT.
dataset activate dataset1.
dataset copy dataset2.
dataset close dataset1.
compute mydate=!maxdate .
execute.
*FILTER OUT THE LAST DATE IN THE DATASET.
DATASET ACTIVATE DataSet2.
FILTER OFF.
USE ALL.
SELECT IF (date < mydate) .
EXECUTE.
save
outfile=!QUOTE(!CONCAT('e:\se\',!league,'\',!venue,'\Data\POST\post_data.sav'))
/ compressed .
!ENDDEFINE.
!newdate level1 level2 .
When I run this locally the syntax works. However, when I run it on the
SPSS Server I get the following error: "Due to contention for the
specified
file, the data has been saved to a file with a different name -
e:\se\level1\level2\data\post\post_data_1.sav"
Can anyone tell me a more efficient way to accomplish this task.
--
View this message in context:
http://spssx-discussion.1045642.n5.nabble.com/Overwriting-a-dataset-tp5080768p5080768.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.
=====================
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
=====================
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
--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Overwriting-a-dataset-tp5080768p5081115.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.
=====================
To manage your subscription to SPSSX-L, send a message to
LISTSERV@LISTSERV.UGA.EDU (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 LISTSERV@LISTSERV.UGA.EDU (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
[text/html]
