Date: Fri, 23 Sep 2011 14:11:09 -0700
Reply-To: David Marso <david.marso@gmail.com>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: David Marso <david.marso@gmail.com>
Subject: Re: the intricacies of using do if-end if and execute
In-Reply-To: <OFAC713F31.D1A5601E-ON87257914.006BF55D-87257914.006C4A74@us.ibm.com>
Content-Type: text/plain; charset=UTF-8
"But you can shift any number of variables in a single data pass."
OK, I can see that now. Current syntax seems a PIA ;-)
Any chance for some later version to provide the following sort of syntactic
flexibility?
SHIFT VALUES varlag_1 TO varlag_10=LAG(singlevar,1,10).
or
SHIFT VALUES lag_1_var1 TO lag_1_var10 = LAG(oldvar1 TO oldvar10,1).
-------
Jon K Peck wrote:
>
> SHIFT VALUES is a procedure, of course. But you can shift any number of
> variables in a single data pass.
> SV was motivated by two requirements
> - ability to handle leads, which the transformation data feed is not set
> up to do.
> - ability to handle unlimited file sizes. CREATE, which can do leads and
> lags, requires the data to be in memory. That is not because of lead or
> lag functionality per se but because of the other functions that CREATE
> supports.
>
> Jon Peck (no "h")
> Senior Software Engineer, IBM
> peck@us.ibm.com
> new phone: 720-342-5621
>
>
>
>
> From: David Marso <david.marso@gmail.com>
> To: SPSSX-L@listserv.uga.edu
> Date: 09/23/2011 01:37 PM
> Subject: Re: [SPSSX-L] the intricacies of using do if-end if and
> execute
> Sent by: "SPSSX(r) Discussion" <SPSSX-L@listserv.uga.edu>
>
>
>
> OTOH: From the FM, SHIFT VALUES results in a data pass!!
> Hence I suspect cannot be done in a LOOP or DO REPEAT and with HUGE data
> sets WILL result in
> unacceptable performance hits.
> For example:
> DO REPEAT X = X1 TO X10 / Y=Y1 TO Y10 / L=1 TO 10 .
> + COMPUTE X=LAG(Xvar,L).
> + COMPUTE Y=LAG(Yvar,L).
> END REPEAT.
> If recast as a set of SHIFT VALUE commands would result in 20 data passes
> and require either a MACRO or 20 separate commands.
> Sacrificing performance for a minor gain in "intuitiveness" is hardly
> laudible ;-)
>
>
> Jon K Peck wrote:
>>
>> Lag is tricky. You might find the SHIFT VALUES behavior to be more
>> intuitive.
>>
>> Jon Peck (no "h")
>> Senior Software Engineer, IBM
>> peck@us.ibm.com
>> new phone: 720-342-5621
>>
>>
>>
>>
>> From: "Pirritano, Matthew" <MPirritano@ochca.com>
>> To: SPSSX-L@LISTSERV.UGA.EDU
>> Date: 09/23/2011 11:49 AM
>> Subject: Re: [SPSSX-L] the intricacies of using do if-end if and
>> execute
>> Sent by: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
>>
>>
>>
>> Thanks David & Rick,
>>
>> It looks like it probably has something to do with the fact that lag is
>> calculated after all other transformations even if it precedes them. So
> my
>> select if is actually happening before the compute using lag that the
>> select if is supposed to be based on.
>>
>> Thanks
>> Matt
>>
>> Matthew Pirritano, Ph.D.
>> Research Analyst IV
>> Medical Services Initiative (MSI)
>> Orange County Health Care Agency
>> (714) 568-5648
>>
>> From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of
>> Rick Oliver
>> Sent: Friday, September 23, 2011 10:25 AM
>> To: SPSSX-L@LISTSERV.UGA.EDU
>> Subject: Re: the intricacies of using do if-end if and execute
>>
>> I don't know what the original problem was. But if lag is an issue, go
>> here:
>>
>> http://publib.boulder.ibm.com/infocenter/spssstat/v20r0m0/index.jsp
>>
>> And then enter "lag function" in the Search field.
>>
>> It may shed some light on the problem.
>>
>>
>>
>> From: David Marso <david.marso@gmail.com>
>> To: SPSSX-L@listserv.uga.edu
>> Date: 09/23/2011 12:12 PM
>> Subject: Re: the intricacies of using do if-end if and execute
>> Sent by: "SPSSX(r) Discussion" <SPSSX-L@listserv.uga.edu>
>>
>>
>>
>>
>> "Whereas if I insert an exe after the do-if I get what I want, a removal
>> of
>> all dups and primary cases, those instances that are a credit debit
> wash.
>> I
>> don’t even see where primary cases are being selected here. Hopefully my
>> story is not too convoluted and someone has followed it."
>>
>> Matt,
>> It has *NOTHING* specifically to do with DO IF and *EVERYTHING* to do
> with
>> the LAG!!!
>> Yeah, your story is way too convoluted. Since I don't use that silly
>> dialog
>> to DEDUP, I have *NO* idea what convoluted syntax you are beating your
>> head
>> against. Example of the data might be useful and the specific desired
>> result. OTOH, whenever selection is done based on variables created by
>> LAG
>> you typically need a procedure to pass the data so things don't get
>> FUBARed.
>> (Might be best to use that data pass to get something informative?
> Perhaps
>> freqs/XTABS on some variables you have just created as reality checks
> ;-)
>> HTH, David
>>
>>
>> --
>> View this message in context:
>>
> http://spssx-discussion.1045642.n5.nabble.com/the-intricacies-of-using-do-if-end-if-and-execute-tp4834252p4834370.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
>>
>
>
> --
> View this message in context:
> http://spssx-discussion.1045642.n5.nabble.com/the-intricacies-of-using-do-if-end-if-and-execute-tp4834252p4834871.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
>
--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/the-intricacies-of-using-do-if-end-if-and-execute-tp4834252p4835219.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
| 
