Date: Wed, 1 Feb 2012 17:25:55 -0800
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: Interesting bit of code ;-)
In-Reply-To: <1328139179991-5449148.post@n5.nabble.com>
Content-Type: text/plain; charset=us-ascii
Yeah, BUT I find the following to be as transparent as glass ;-).
ONE compute per case rather than 50 and NO logical comparison required.
NUMERIC state_01 TO state_50 (F1).
RECODE state_01 TO state_50 (ELSE=0).
VECTOR state_dummy=state_01 TO state_50.
COMPUTE state_dummy(state)=1.
"p.s. - I'd never noticed that one can use PRINT like that on END REPEAT.
Thanks for educating me (once again)."
Glad to be of service!
--
** You can also use:
END REPEAT NOPRINT (but WTF? there for the sake of completeness???).
--
Bruce Weaver wrote
>
> You're right about efficiency. But unless one is working with a HUGE data
> file, the difference will likely be imperceptible (to the human eye, at
> least). And in that case, transparency should trump efficiency.
>
> I feel like I'm stealing material from Art Kendall here. ;-)
>
> p.s. - I'd never noticed that one can use PRINT like that on END REPEAT.
> Thanks for educating me (once again).
>
>
>
> David Marso wrote
>>
>> OTOH It is less efficient than the VECTOR approach.
>> N computes rather than 1.
>> I'm not sure about the RECODE WRT processing efficiency. My point was
>> the interesting way that RECODE allows multiple vars to be created from
>> one variable and the ability to subsequently recode these new variables
>> in the single recode statement.
>> --
>> Remember DO REPEAT cycles through the entire list of stand in
>> 'variables'.
>> --
>> NUMERIC dx1 TO dx4 (F1).
>>
>> DO REPEAT dx = dx1 to dx4 / # = 1 to 4 .
>> - COMPUTE dx = g4 EQ #.
>> END REPEAT PRINT.
>>
>> 18 0 +COMPUTE DX1 = G4 EQ 1
>> 19 0 +COMPUTE DX2 = G4 EQ 2
>> 20 0 +COMPUTE DX3 = G4 EQ 3
>> 21 0 +COMPUTE DX4 = G4 EQ 4
>>
>> LIST.
>>
>> Bruce Weaver wrote
>>>
>>> I like DO-REPEAT for this task. It's very transparent, I think, and not
>>> significantly more likely to cause RSI than the other methods you show .
>>> ;-)
>>>
>>> DATA LIST FREE / g4 (F1).
>>> BEGIN DATA
>>> 1 2 3 4
>>> END DATA.
>>>
>>> NUMERIC dx1 TO dx4 (F1).
>>> DO REPEAT dx = dx1 to dx4 / # = 1 to 4 .
>>> - COMPUTE dx = g4 EQ #.
>>> END REPEAT.
>>> LIST.
>>>
>>>
>>>
>>> David Marso wrote
>>>>
>>>> DUMMY Variables...
>>>> --------------------
>>>> *G4 exists in the data file and has integer values between 1 and 4.
>>>> **COMMENTS??**.
>>>> RECODE G4 (1=1) INTO DX1
>>>> / G4 (2=1) INTO DX2
>>>> / G4 (3=1) INTO DX3
>>>> / DX1 DX2 DX3 (MISSING=0).
>>>>
>>>> OTOH: The following is more concise with large number of groups ;-)
>>>>
>>>> NUMERIC DX1 TO DX4 (F1).
>>>> RECODE DX1 TO DX4 (ELSE=0).
>>>> VECTOR DX=DX1 TO DX4.
>>>> COMPUTE DX(G4)=1.
>>>>
>>>> The following IMNSHO is abysmal.
>>>> DO IF G4=1.
>>>> + COMPUTE DX1=1.
>>>> ELSE IF G4=2.
>>>> + COMPUTE DX2=1.
>>>> ELSE IF G4=3.
>>>> + COMPUTE DX3=1.
>>>> END IF.
>>>> RECODE DX1 TO DX4 (MISSING=0)(ELSE=COPY).
>>>>
>>>
>>
>
--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Interesting-bit-of-code-tp5448698p5449349.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
