Date: Tue, 16 Dec 1997 16:40:49 -0500
Reply-To: Victor Kamensky <kamensky@AECOM.YU.EDU>
Sender: "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From: Victor Kamensky <kamensky@AECOM.YU.EDU>
Subject: Re: Strange results
Content-Type: text/plain; charset="us-ascii"
Thanks to Tim Berryhill-the problem is solved.
The following data step should be added:
data d1 ; set d1;
if qnum05=9 then quarter=4;
For the second run results were correct,but it
was due to the data: all the records
at the end of the nonsorted data set have qnum05=9.
Victor Kamensky
At 04:02 PM 12/15/97 PST, you wrote:
>Victor, I notice that there are no variables in common between D1 and D2, and
>that they are concatenated (first all observations of D1 are processed,
then all
>observations of D2). In this case, SAS creates the program data vector for D1
>and D3 with all variables from each dataset (the log shows four variables).
>Before processing ANY data, all four variables are missing. Once the first
>observation from D1 is read, ID and QNUM05 have values. Get ready:
>
>!! QUARTER does not have a value until either 1) you read the first observation
>from D2 or 2) an observation from D1 has a QNUM05 of 9. Once some D1
>observation with a QNUM05 value of 9 sets QUARTER to 4, QUARTER remains 4 until
>an observation from D2 changes that value.
>
>To test this notion, you might insert a few put statements into the final data
>step (both before and after the SET statement). I think you will find the
>results were never correct.
>
>Tim Berryhill - Contract Programmer and General Wizard
>TWB2@PGE.COM or http://www.aartwolf.com/twb.html
>Frequently at Pacific Gas & Electric Co., San Francisco
>The correlation coefficient between their views and
>my postings is slightly less than 0
>----------------------[Reply - Original Message]----------------------
>
>Sent by:"Victor Kamensky" <kamensky@AECOM.YU.EDU>
> Dear SAS-L-ers!
>Please take a look at the following log.
>I run the same program twice.
>The difference between first and second run is that
>proc sort is taken out.
>First run gives wrong results,second-correct.
>Data base d2 is empty.
>(Another way of getting correct results
> is to drop QUARTER from d2,although d2 is empty).
>It does not make any sense.
>Any ideas?
>Sincerely
>Victor Kamensky
>Programmer
>Albert Einstein College of Medicine
>
>
>18313 *dtt.sas;
>18314 *FIRST RUN-WRONG RESULTS;
>18315 PROC DBF DB4=d1 OUT=d1(keep=id qnum05);RUN;
>
>NOTE: 937 observations written to the output SAS data set.
>NOTE: The PROCEDURE DBF used 0.22 seconds.
>
>
>18316 PROC DBF DB4=d2 OUT=d2(keep=acrostic quarter);RUN;
>
>NOTE: 0 observations written to the output SAS data set.
>NOTE: The PROCEDURE DBF used 0.05 seconds.
>
>
>18317 proc sort data=d1; by id; run;
>
>NOTE: The data set WORK.D1 has 937 observations and 2 variables.
>NOTE: The PROCEDURE SORT used 0.11 seconds.
>
>
>18318 data d1 d3;
>18319 set d1 d2(keep=acrostic quarter);
>18320 if qnum05=9 then quarter=4;
>18321 if quarter =4 then output d3;
>18322 else output d1;
>18323 run;
>
>NOTE: The data set WORK.D1 has 359 observations and 4 variables.
>NOTE: The data set WORK.D3 has 578 observations and 4 variables.
>NOTE: The DATA statement used 0.55 seconds.
>
>
>18324
>18325 *SECOND RUN-NO SORTING-CORRECT RESULT;
>18326 PROC DBF DB4=d1 OUT=d1(keep=id qnum05);RUN;
>
>NOTE: 937 observations written to the output SAS data set.
>NOTE: The PROCEDURE DBF used 0.16 seconds.
>
>
>18327 PROC DBF DB4=d2 OUT=d2(keep=acrostic quarter);RUN;
>
>NOTE: 0 observations written to the output SAS data set.
>NOTE: The PROCEDURE DBF used 0.05 seconds.
>
>
>18328 *proc sort data=d1; *by id; run;
>18329 data d1 d3;
>18330 set d1 d2(keep=acrostic quarter);
>18331 if qnum05=9 then quarter=4;
>18332 if quarter =4 then output d3;
>18333 else output d1;
>18334 run;
>
>NOTE: The data set WORK.D1 has 704 observations and 4 variables.
>NOTE: The data set WORK.D3 has 233 observations and 4 variables.
>NOTE: The DATA statement used 0.44 seconds.
>
>=====================================================================
>
>
| 
