struggling with OMS
Locked
 
|
Dear colleagues,
I am having a difficult time with OMS. I am new to this feature of SPSS. I scanned through the Syntax Reference guide and the list serve archives but without success. My issue. I am conducting a Friedman test and want to export the Ranks table to a *.sav file such that the variable names in the aforementioned table will be the variable names in the new data file and the mean ranks will be the values. Ideally I want to export more than one table such that each table will be a case in the new data file. Currently I used this syntax for just one table: OMS /SELECT TABLES /IF COMMANDS=['NPar Tests'] SUBTYPES=['Friedman Ranks'] LABELS =['Mean Rank'] /DESTINATION FORMAT=SAV OUTFILE='C:\myfolder\myfile.sav' /COLUMNS SEQUENCE=[RALL CALL LALL]. OMSEND. All this syntax provides is a data file with "Command_", "Subtype_" and "Label_" as variables but no data. I assume that the OMS command reads the output (SPSS Viewer) currently open - that is where the results of the Friedman test are located. The "[RALL CALL LALL]" puts all dimensions in a single row. That might be a problem since the Ranks table has two columns: one with the variables and the other with the ranks. I tried using DIMNAMES but without success. Perhaps I need to transpose rows and columns in the original table in Viewer? Any suggestions would be much appreciated. Thanks in advance. Best regards, Dominic Dominic Lusinchi Far West Research Statistical Consulting San Francisco, California 415-664-3032 [hidden email] |
|
|
Hi Domonic,
OMS does not work on existing output like OUTPUT EXPORT. It's a new output stream. You must put your table-producing procedure between the OMS and OMSEND commands. Good luck, Pierre Tourigny On 2/20/2012 3:04 AM, Dominic Lusinchi wrote: Dear colleagues, |
|
|
As Pierre says, OMS listens for output
generated after you start it. It captures it upstream from the Viewer.
You can pivot tables in the OMS capture, but OMS provides no control
over variable naming.
If you need to capture the output exactly in the Viewer format, it would require custom scripting in Python or Basic. Jon Peck (no "h") aka Kim Senior Software Engineer, IBM [hidden email] new phone: 720-342-5621 From: Pierre Tourigny <[hidden email]> To: [hidden email] Date: 02/20/2012 04:48 AM Subject: Re: [SPSSX-L] struggling with OMS Sent by: "SPSSX(r) Discussion" <[hidden email]> Hi Domonic, OMS does not work on existing output like OUTPUT EXPORT. It's a new output stream. You must put your table-producing procedure between the OMS and OMSEND commands. Good luck, Pierre Tourigny On 2/20/2012 3:04 AM, Dominic Lusinchi wrote: Dear colleagues, I am having a difficult time with OMS. I am new to this feature of SPSS. I scanned through the Syntax Reference guide and the list serve archives but without success. My issue. I am conducting a Friedman test and want to export the Ranks table to a *.sav file such that the variable names in the aforementioned table will be the variable names in the new data file and the mean ranks will be the values. Ideally I want to export more than one table such that each table will be a case in the new data file. Currently I used this syntax for just one table: OMS /SELECT TABLES /IF COMMANDS=['NPar Tests'] SUBTYPES=['Friedman Ranks'] LABELS =['Mean Rank'] /DESTINATION FORMAT=SAV OUTFILE='C:\myfolder\myfile.sav' /COLUMNS SEQUENCE=[RALL CALL LALL]. OMSEND. All this syntax provides is a data file with "Command_", "Subtype_" and "Label_" as variables but no data. I assume that the OMS command reads the output (SPSS Viewer) currently open - that is where the results of the Friedman test are located. The "[RALL CALL LALL]" puts all dimensions in a single row. That might be a problem since the Ranks table has two columns: one with the variables and the other with the ranks. I tried using DIMNAMES but without success. Perhaps I need to transpose rows and columns in the original table in Viewer? Any suggestions would be much appreciated. Thanks in advance. Best regards, Dominic Dominic Lusinchi Far West Research Statistical Consulting San Francisco, California 415-664-3032 dominic@... |
|
|
In reply to this post by Dominic Lusinchi
Hi Dominic, I played around with the problem you posed and I came up with the start of a possible solution: DATA LIST LIST / a1 a2 a3 . BEGIN DATA 3.0 4.0 7.0 6.0 5.0 8.0 3.0 3.0 6.0 2.0 3.0 6.0 2.0 3.0 6.0 END DATA. OMS /SELECT TABLES /IF COMMANDS = ['NPar Tests']
SUBTYPES = [ 'Friedman Ranks' ] /DESTINATION FORMAT =SAV OUTFILE ='C:\myfolder\myfile.sav' . NPAR TESTS /FRIEDMAN=a1 a2 a3. NPAR TESTS /FRIEDMAN=a1 a2 a3. OMSEND. GET FILE = 'C:\myfolder\myfile.sav' . RECODE Var1 ('a1' = 1)('a2' = 2) ('a3' = 3) INTO varA . EXECUTE. DELETE VARIABLES Command_ Subtype_ Label_ Var1. EXECUTE, * Insert a case variable to indicate rows that go together This probably can be automated.
CASESTOVARS /ID = case /INDEX = VarA . LIST. I hope this gives you some ideas.
Good luck,
Stephen Salbod, Pace University, NYC From: SPSSX(r) Discussion [mailto:[hidden email]]
On Behalf Of Dominic Lusinchi Dear colleagues, |
|
|
Thank you, Stephen, for your efforts. If I read your syntax correctly, it appears
that OMS in itself cannot do what I am looking for: i.e. to have a1, a2, and a2
as variables and mean ranks as values for each variable. I thought the
subcommand COLUMNS was used for that purpose (?). Also, it would seem that my assumption (“the
OMS command reads the output (SPSS Viewer) currently open”) is incorrect.
The procedure that generates the output of interest (in this case Friedman’s
test) has to be embedded within the OMS-OMSEND structure (?). In other word,
one can’t just run the procedure and obtain an output and then run OMS
with that output open and expect OMS to read that output (?). Good of you to take the time. Best regards, Dominic Dominic Lusinchi Statistical Consulting T/F 415-664-3032 Instructor, UC Berkeley Extension: From: Salbod, Mr.
Stephen [mailto:[hidden email]] Hi Dominic, I played around with
the problem you posed and I came up with the start of a possible solution: DATA LIST LIST / a1
a2 a3 . BEGIN DATA 3.0
4.0 7.0 6.0
5.0 8.0 3.0
3.0 6.0 2.0
3.0 6.0 2.0
3.0 6.0 END DATA. OMS /SELECT
TABLES /IF COMMANDS
= ['NPar Tests'] SUBTYPES
= [ 'Friedman Ranks' ]
/DESTINATION FORMAT =SAV
OUTFILE ='C:\myfolder\myfile.sav' . NPAR TESTS
/FRIEDMAN=a1 a2 a3. NPAR TESTS
/FRIEDMAN=a1 a2 a3. OMSEND. GET FILE =
'C:\myfolder\myfile.sav' . RECODE Var1 ('a1' =
1)('a2' = 2) ('a3' = 3) INTO varA . EXECUTE. DELETE VARIABLES
Command_ Subtype_ Label_ Var1. EXECUTE, * Insert a case
variable to indicate rows that go together This probably can be automated. CASESTOVARS /ID = case /INDEX = VarA . LIST. I hope this gives
you some ideas. Good luck, Stephen Salbod, From: SPSSX(r)
Discussion [mailto:[hidden email]] On
Behalf Of Dominic Lusinchi Dear colleagues, |
|
|
In reply to this post by Dominic Lusinchi
Thanks, Jon (and Pierre). Based on Stephen’s
response I understand now the “listening” part: the procedure of
interest must be embedded into OMS. Using Stephen’s data, I get what I
want with this syntax: OMS /SELECT TABLES /IF COMMANDS=['NPar Tests']
SUBTYPES=['Friedman Ranks'] /DESTINATION FORMAT=SAV OUTFILE='C:\myfolder\myfile.sav' /COLUMNS SEQUENCE=[RALL CALL LALL]. NPAR TESTS /FRIEDMAN=a1 a2 a3. OMSEND. Thank you all. Dominic Lusinchi Statistical Consulting T/F 415-664-3032 Instructor, UC Berkeley Extension: From: Jon K Peck
[mailto:[hidden email]] As
|
Automatic reply: struggling with OMS
|
|
Out of office 2/21-2/22. |
| Free forum by Nabble | Edit this page |