Re: Basic do loop question

Mark ,

I am going to tweak your code out using V9, hope ya dont mind:


Data Need ( Drop = I Old_Nbr_Mc ) ;
Set Sample ( Rename = ( Nbr_Mc = Old_Nbr_Mc ) ) ;

Do I = 1 To ( CountC( Old_Nbr_Mc , ',' ) + 1 ) ;
Nbr_Mc = Scan( Old_Nbr_Mc , I , ',' ) ;
Output ;
End ;

Run ;



Toby Dunn

Quickly, bring me a beaker of wine, so that I may wet my mind and say
something clever.
Aristophanes

Wise people, even though all laws were abolished, would still lead the same
life.
Aristophanes

You should not decide until you have heard what both have to say.
Aristophanes






From: "Terjeson, Mark" <Mterjeson@RUSSELL.COM>
Reply-To: "Terjeson, Mark" <Mterjeson@RUSSELL.COM>
To: SAS-L@LISTSERV.UGA.EDU
Subject: Re: Basic do loop question
Date: Wed, 22 Nov 2006 08:03:14 -0800

Hi Jeri,

Providing a small sample of what you have,
and what you want, is perfect for getting
more concise and applicable answers. What
you desire really does not need the number
of observations, but the number of mc###
elements within one variable only. The
iterative nature of the DATA step will
automatically iterate through all the
incoming observations with the SET statement.
All the lines of code between the DATA and
RUN statements will execute on each obs.
Without an OUTPUT statement there is one
implied OUTPUT at the RUN statement. If we
add an OUTPUT statement anywhere in our
datastep then the automatic one at the RUN
statement is disabled and you basically take
manual control of the outputs whereever your
one or more OUTPUT statements occur. With
this in mind then all we have to do is output
whatever is in the PDV (Program Data Vector)
[the input/output variable buffer] for as
many times as we find mc### in the nbr_mc
variable. Therefore we do a couple things
to accomplish that. First, we are going to
need to use the nbr_mc variable for our
output containing only one mc### element at
a time, so we need to initially upon each
observation save the whole string off to
a temporary character variable(stmp) so
that we keep all of the mc###'s. At this
point we could count commas and just add 1
to the result to reach the number of mc###
arguments. Or at the same time if we add one
extra comma and not add 1 to get the number
of args. Both achieve the same result but
just in case you have an empty string we can
avoid a little SAS nuance that you always get
1 from the LENGTH() function, which this scheme
avoids. Then a mere DO loop from 1 to the
number of commas and the SCAN function will
grab string chunks inbetween whatever delimiter
you specify. Effectively chopping out each
mc### and placing each in the nbr_mc variable.
An OUTPUT at that point gets you your result
set as desired. The TRIM() function is used
to truncate the trailing spaces of the nbr_mc
variable before concatenating on the extra
comma, this way the added comma is right against
the end of the text. The DROP will throw away
the temporary variables so that they do not end
up in the outgoing dataset.


data sample;
input id nbr_mc $24. detail;
cards;
12345 mc134,mc345,mc324 1.25
67891 mc333 3.11
56547 mc324,mc456 3.25
;
run;


data result(drop=i comma_count stmp);
set sample;
stmp = trim(nbr_mc)||',';
comma_count = length(stmp)-length(compress(stmp,','));
do i = 1 to comma_count;
nbr_mc = scan(stmp,i,',');
output;
end;
run;






Hope this is helpful.


Mark Terjeson
Senior Programmer Analyst, IM&R
Russell Investment Group


Russell
Global Leaders in Multi-Manager Investing






-----Original Message-----
From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of
Jeri Ji
Sent: Wednesday, November 22, 2006 6:46 AM
To: SAS-L@LISTSERV.UGA.EDU
Subject: Re: Basic do loop question

Thank you very much, Kevin and Mark. Both of you answered my question. I
will try the code that you guys suggested.

What I need to do is that I have a dataset look like the following:

id nbr_mc detail
12345 mc134,mc345,mc324 1.25
67891 mc333 3.11
56547 mc324,mc456 3.25
... . .
... . .

And I need the final result to look like this:
id nbr_mc detail
12345 mc134 1.25
12345 mc345 1.25
12345 mc324 1.25
67891 mc333 3.11
56547 mc324 3.25
56547 mc456 3.25

I am a new sas user and never used do loop before. I think maybe I can
use a outer do loop loop from the 1st record to the end, and then an
inner loop to break up the nbr_mc. The thing is first the number of rows
in the dataset is not fixed. Second, the number of mc numbers in that
column is not fixed either. I have the idea, but didn't figure out how
to code it yet.

Thank you very much.



Jeri



Kevin Roland Viel <kviel@EMORY.EDU>
Sent by: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
11/21/2006 06:58 PM
Please respond to
Kevin Roland Viel <kviel@EMORY.EDU>


To
SAS-L@LISTSERV.UGA.EDU
cc

Subject
Re: Basic do loop question






On Tue, 21 Nov 2006, Jeri Ji wrote:

> I am going to use do loop the first time in sas and have a basic
question.
> what should I do if I want the do loop start from the 1st row in the
> dataset and exit after the last row (the number of rows are not fixed
for
> each run)?
>
> do i=1 to ?

Jeri,

SAS provides this with a minimal amount of coding:

data two ;
set one ;
run ;

Of course, it is implicit. You want something more:

data two ;
do _n_ = 1 by 1 until ( end ) ;
set one end = end ;
end ;
run ;

This is a DOW-loop, or a Whitlock Do-Loop. Note that the default
OUTPUT statement is no longer in effect.

You could also use:

data two ;
do _n_ = 1 to nobs ;
set one nobs = nobs ;
end ;
STOP ;
run ;

If you describe your goal and suggest a design, we could probably
offer more tailored advice...

HTH,

Kevin

PS You may also want to look at the POINT= option to the SET statement.

Kevin Viel
PhD Candidate
Department of Epidemiology
Rollins School of Public Health
Emory University
Atlanta, GA 30322

__________________________________________________ _______________
Talk now to your Hotmail contacts with Windows Live Messenger.
http://clk.atdmt.com/MSN/go/msnnkwme...enger/overview