Re: weblog to normal form

>It looks like the goal is to place each field in the form into a SAS
variable. That seems pretty natural to me.
Natural for some, but not 'normal form' if any of the column variables
have functional dependencies not related to key dependencies ... And, I
fear even to ask, what is happening to key values required to identify a
Web submission?

>at least 1 post: Why would you want to change it?
>Followed by some remarks on bad design of the electrical wiring in the
room. (From another thread ...)
I plead guilty as charged. It seems to me that SAS programmers tend to
spend far too much time and effort trying to turn the clock back to the
flatfile era.
Sig
-----Original Message-----
From: owner-sas-l@listserv.uga.edu [mailtowner-sas-l@listserv.uga.edu]
On Behalf Of Howard Schreier <hs AT dc-sug DOT org>
Sent: Tuesday, April 11, 2006 11:22 PM
To: SAS-L@LISTSERV.UGA.EDU
Subject: Re: weblog to normal form


On Tue, 11 Apr 2006 19:38:41 -0700, David L Cassell
<davidlcassell@MSN.COM>
wrote:

>mysasbox@GMAIL.COM wrote:
>>I have weblog in the following form:
>>
>>data weblog;
>>input name $ value $;
>>datalines;
>>var1 abc
>>var2 efg
>>var3 hij
>>var1 klm
>>var2 nop
>>var3 qrs
>>var1 tuv
>>;
>>run;
>>
>>
>>I need this normal form:
>>
>>var1 var2 var3
>>
>>abc efg hij
>>klm nop qrs
>>tuv
>>
>>I haven't been able to use proc transpose to do this transformation.
>>Would someone be so kind as to point me to a datastep solution.
>
>Mike:
>
>[1] As Alan pointed out, you have something that isn't a weblog. What

>is it?

Looks to me like the capture from a series of web form submissions.

>
>[2] I wouldn't call this a 'normal' form. An abnormal form, yes. :-)

>I sort of wonder why you want this second form, since it looks like it
>will make your work a lot harder.

It looks like the goal is to place each field in the form into a SAS
variable. That seems pretty natural to me.

>
>[3] So I won't point out that a quick sort before your transpose would

>probably solve your problem, since I think that may be leading you down

>the wrong path.

Toby actually coded that, but it yields the transpose of the desired
result.

>
>[4] So what are you trying to do, and why do you think this transposed

>format is going to help you? If all you're doing is data display,
>consider using PROC TABULATE or PROC REPORT instead. If the second
>variable is really something from a weblog, then it may be way too long

>to conveniently display as an 'across' variable in output.
>
>HTH,
>David
>--
>David L. Cassell
>mathematical statistician
>Design Pathways
>3115 NW Norwood Pl.
>Corvallis OR 97330
>
>_________________________________________________ ________________
>Express yourself instantly with MSN Messenger! Download today - it's
>FREE! http://messenger.msn.click-url.com/g...ave/direct/01/