OMS output structure improvements

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

OMS output structure improvements

Maguin, Eugene
I don't know whether other folks use OMS and like the output file structure but I believe that some improvements could be made in the output file structure, particularly, for the Means and Regression commands. My context for this is running certain output elements from a (short) stack of Means or Regression commands into files, one for each element type.

I would like to see a column for command number (i.e., first Means command, second Means command, etc) and another column for command dependent variable. I notice that the Regression coefficients file combines independent variable name and variable label in the same field. There'd be better post-processing flexibility if those elements were put in separate fields. On the other hand (and I may be misremembering) but I believe that for the Means command the labels for the values of the independent variable are not reported, which means they have to be added by hand.

As it is, OMS does not help too much in table building; it's easier to export specific output element types to different RTF files and do post-processing/table building in word--which is unfortunate, I think.

Gene Maguin

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (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
Reply | Threaded
Open this post in threaded view
|

Re: OMS output structure improvements

Jon K Peck
OMS is agnostic to the type of table it is working on.  That's why it can work on every pivot table Statistics produces.  It does give you a table number.  You might find it helpful to have OMS pivot some of the output depending on what you want to do with this, but OMS knows nothing about specific procedures, so it isn't amenable to table-type customization.

As for the labelling in the generated dataset, OMS follows your Edit>Options>Output Labels setting.  You apparently have that set to both values and labels.  Change it to one or the other, and OMS will follow suit.

Of course, it is possible to collect all the pieces you want and build a table using Python programmability.  OMS can save its results to the XMLworkspace, which Python can read and use to build new output.  This is, no doubt, an advanced skill, but for purposes of building production jobs, it is will worth learning.

Regards,

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
[hidden email]
new phone: 720-342-5621




From:        "Maguin, Eugene" <[hidden email]>
To:        [hidden email],
Date:        01/03/2013 09:43 AM
Subject:        [SPSSX-L] OMS output structure improvements
Sent by:        "SPSSX(r) Discussion" <[hidden email]>




I don't know whether other folks use OMS and like the output file structure but I believe that some improvements could be made in the output file structure, particularly, for the Means and Regression commands. My context for this is running certain output elements from a (short) stack of Means or Regression commands into files, one for each element type.

I would like to see a column for command number (i.e., first Means command, second Means command, etc) and another column for command dependent variable. I notice that the Regression coefficients file combines independent variable name and variable label in the same field. There'd be better post-processing flexibility if those elements were put in separate fields. On the other hand (and I may be misremembering) but I believe that for the Means command the labels for the values of the independent variable are not reported, which means they have to be added by hand.

As it is, OMS does not help too much in table building; it's easier to export specific output element types to different RTF files and do post-processing/table building in word--which is unfortunate, I think.

Gene Maguin

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (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