Thanks for all your help on helping me solve my problems. Your suggestions were all great but there was one that worked on my problem best and appears below.

In this case, Art Kendall led me to my solution since I provided him with more information about my problem. I was using a Get Capture command to retrieve ODBC and then selected variables I was interested. Art suggested that I use the CACHE statement after my Get Capture statement when retrieving over 1 million cases from ODBC. Of course, as Art suggested and I knew I could modify my sql statements to include the selection criteria (which would be the best way around it). Anyhow, his suggested of placing Cache after the Get Capture cmd saved me about 4 to 10 minutes.

So the command sequence in the syntax looks something like this

Get Capture ODBC Cache. Select if Save outfile

Thanks, Manny

-----Original Message----- From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of Oliver, Richard Sent: Tuesday, May 11, 2004 4:23 PM To: SPSSX-L@LISTSERV.UGA.EDU Subject: Re: large files with SPSS: how does spss handle data

select if... cache. execute. *or some other command that reads the data

should do the job.

-----Original Message----- From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU]On Behalf Of Art Kendall Sent: Tuesday, May 11, 2004 3:18 PM To: SPSSX-L@LISTSERV.UGA.EDU Subject: Re: large files with SPSS: how does spss handle data

Are you reading in an SPSS system file or are you using SQL?

Does the data reside on the your local machine or on a remote server?

Please post the syntax you are using to select the 1600 cases.

In earlier versions I think you would only use cache immediately after GETting a file from a remote server.

After the first procedure (e.g. execute) the data view should show only 1600 cases if you used SELECT IF.

Older versions of SPSS were smart enough to not create the scratch file unless there was more than one procedure.

try something like this. get file file= ... select if . . . save outfile= 'd:\project\selected 1600.sav'. new file. get file= 'd:\project\selected 1600.sav'.

Art Art@DrKendall.org Social Research Consultants University Park, MD USA (301) 864-5570

Manny Straehle wrote:

>Is this available in 12? If not is their a way to set cache settings globally >in 12? If not where do I place the CACHE command after every select >statement before the execute. I never used it before. > >Thanks, >Manny > > >Manfred M. Straehle >Scoring Services >mstraehle@nbme.org >215-590-9873 > > > > >-----Original Message----- >From: Bob Walker [mailto:spssx-l@surveys-forecasts.com] >Sent: Tuesday, May 11, 2004 3:48 PM >To: Manny Straehle; SPSSX-L@LISTSERV.UGA.EDU >Subject: RE: large files with SPSS: how does spss handle data > >Another thought... earlier versions of SPSS have a workspace setting that >can be adjusted in the options dialog. It can also be implemented via syntax >via SET WORKSPACE. If you're using on older version, this may help to >improve overall performance. Regards, > >Bob Walker >Surveys & Forecasts, LLC >www.surveys-forecasts.com > > >-----Original Message----- >From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of >Manny Straehle >Sent: Tuesday, May 11, 2004 3:34 PM >To: SPSSX-L@LISTSERV.UGA.EDU >Subject: large files with SPSS: how does spss handle data > >Quick Question: > >Does anybody know when dealing with large files (over 1mil cases) how does >SPSS use hidden temporary files or memory to transform data? For instance, I >have noticed that when I have many cases and wish to use a select statement >SPSS will slow down not only for the select statement with executable (which >is expected) but even after it has executed the select statement.0 > >Why is this the case even after I selected fewer cases (1600 from over >1000000)? It seems that SPSS still retains the many (1000000) cases either >in a file or in RAM somewhere after you executed the selection of 1600 >cases. >If this is the case can I somehow change this and what will be the >consequences if I can change it? > >Anybody have any idea? Jon? > >Thanks, >Manny > > > >-----Original Message----- >From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of >Peck, Jon >Sent: Tuesday, May 04, 2004 2:44 PM >To: SPSSX-L@LISTSERV.UGA.EDU >Subject: Re: Delete variable > >In SPSS 12, you can just say >DELETE VARIABLES list-of-variables. > >The list honors the TO convention, but you can't delete ALL your variables >this way. > >And you can't use this command in the middle of a set of transformations. > >Regards, >Jon Peck > >-----Original Message----- >From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU]On Behalf Of >Marks, Jim >Sent: Tuesday, May 04, 2004 1:18 PM >To: SPSSX-L@LISTSERV.UGA.EDU >Subject: Re: [SPSSX-L] Delete variable > > >Jakob: > >ADD FILES FILE = * /DROP var1. >EXECUTE. > >jim > >-----Original Message----- >From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU]On Behalf Of >Jakob Fiellau >Sent: Tuesday, May 04, 2004 1:15 PM >To: SPSSX-L@LISTSERV.UGA.EDU >Subject: Delete variable > > >Hi > >Can anyone give me the syntax for deleting a varaible - e.g DELETE VAR1 > >Thanks >Jakob Fiellau > > >