See below

Jon Peck (no "h") aka Kim Senior Software Engineer, IBM peck@us.ibm.com new phone: 720-342-5621

From: "Bibel, Daniel (POL)" <daniel.bibel@state.ma.us> To: SPSSX-L@listserv.uga.edu Date: 03/13/2012 08:13 AM Subject: [SPSSX-L] Question about spssaux2 and V20 Sent by: "SPSSX(r) Discussion" <SPSSX-L@listserv.uga.edu>

I've just installed a trial version of V20 and come across a piece of code that doesn't work as it did in V18.

I have a file with monthly data from agencies. I want to split the files by variable ('agency') and save the output into separate Excel files. I've installed the Python essentials and copied the spssaux2.py file from the c:\python26 directory into the c:\python27 directory.

>>>I would always recommend getting the latest version of a supplementary file from the SPSS Community website rather than copying an old version, although in this case, it makes the behavior comparison simpler.

The code that worked previously is:

>begin program. >import spss, spssaux2 >cmd = "SAVE TRANSLATE OUTFILE='U:/cru_data/SummaryReports/monthly0011/%(splitvalue)s.xls' /TYPE=XLS /VERSION=8 /replace /fieldnames /replace "

>spssaux2.generalizedSplit('Agency', cmd)

>end program.

The output I get now is like this:

begin program. import spss, spssaux2 cmd = "SAVE TRANSLATE OUTFILE='U:/cru_data/SummaryReports/monthly0011/%(splitvalue)s.xls' /TYPE=XLS /VERSION=8 /replace /fieldnames /replace " spssaux2.generalizedSplit('Agency', cmd) end program.

SPLIT: Agency = Abington >Error # 9016 in column 28. Text: U:/cru_data/SummaryReports/monthly0011/"Abington".xls >The file specification is invalid, and the output file cannot be opened. >Execution of this command stops.

Any suggestions greatly appreciated! >>>I don't think this code could have worked in V18 with a string variable. The split values for strings are quoted, and the quotes are causing the problem in the output file name. In fact, I tried a similar example with V18 and got the same failure.

You might be interested in the SPSSINC SPLIT DATASET extension command, which provides a more general way to do this type of operation and doesn't require you to write any Python code.

generalizedsplit could be modified to allow a reference to the unquoted value, but SPLIT DATASET provides a more powerful tool.

HTH, Jon Peck Dan Bibel Massachusetts State Police