Date:         Tue, 15 Apr 2003 05:11:11 -0700
Reply-To:     Rune Runnestoe <rune@FASTLANE.NO>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Rune Runnestoe <rune@FASTLANE.NO>
Organization: http://groups.google.com/
Subject:      Evaluating the sorting order in one and two columns
Content-Type: text/plain; charset=ISO-8859-1

/*
Kan someone help med with the code where I have written it as
pseudocode ?
*/

/*Making the table SAK, and filling it up with data*/
proc sql;
create table test.sak
(Saksnr varchar(8));
insert into test.sak
values ('2000/122')
values ('2000/123')
values ('1999/981')
values ('2001/585')
values ('2001/586')
values ('2001/584')
values ('2001/587')
values ('2000/124');

/*I am going to identify rows that are sorted wrong and rows that are
sorted right in SAK. The sorting is demanded ascending.*/

data test.SortSak_right test.SortSak_wrong;
set test.sak;
/*
Demands for the sorting of test.sak:
For each row, compare the values of Saksnr with the value of all the
rows beneath it.
If any of the rows beneath is lower, then output test.SortSak_wrong.
If none of the values in the beneath rows are lower, then output
test.SortSak_right.

if input(substr(test.Saksnr,1,4)) on this row is greater than
   input(substr(test.Saksnr,1,4)) on the rows beneath, then
        output test.SortSak_feil;
else if input(substr(test.Saksnr,1,4)) on this row is less than
   input(substr(test.Saksnr,1,4)) on rows beneath then
        output test.SortSak_rett;
else if input(substr(test.Saksnr,1,4)) on this row is like
    input(substr(test.Saksnr,1,4)) på rader nedanfor then
        do;
          if substr(test.Saksnr,6,3) on this row is greater than
             substr(test.Saksnr,6,3) on other rows then
                        output test.SortSak_wrong;
          else if substr(test.Saksnr,6,3) on this row is less than
             substr(test.Saksnr,6,3) on other rows then
                        output test.SortSak_rett;
        end;
*/
run;

/*Making the table DOK, and filling it up with data*/
proc sql;
create table test.dok
(Saksnr varchar(8),
Doknr   num(3));
insert into test.dok
values ('2001/585',1)
values ('2000/123',1)
values ('2000/123',3)
values ('2000/123',2)
values ('2000/122',1)
values ('2000/122',2)
values ('1999/981',1)
values ('2001/586',2)
values ('2001/586',1)
values ('2001/586',3);


/*I am going to find the rows that are wrong and right sorted in DOK.
The sorting is demanded to be ascending, first for the column Saksnr,
then for the column Doknr.
*/
data test.SortDok_right SortDok_wrong;
set test.dok;
/*
If input(substr(Saksnr,1,4)) on this row is greater than
   input(substr(Saksnr,1,4)) on the rows beneath then
        output test.SortDok_wrong;
else if input(substr(Saksnr,1,4)) on this row is less than
   input(substr(Saksnr,1,4)) on the rows beneath then
        output test.SortDok_right;
else if input(substr(Saksnr,1,4)) on this row is like
   input(substr(Saksnr,1,4)) on the rows beneath then
    do;
       If Doknr on this row is greater than Doknr on other rows with
the
             same Saksnr then
                        output test.SortDok_wrong;
       else if Doknr on this row is less than Doknr on other rows with
the
             same Saksnr then
                        output test.SortDok_right;
    end;

Is there a smarter or more simple programming logic for this task ?
Can a while-loop or a for-loop be more suitable  ? */

run;


/*Outprint from the table SAK*/
proc sql;
title "Review of wrongly sortet rows in SAK";
select *
from test.SortSak_wrong;

/*Outprint from the table DOK*/
proc sql;
title "Review of wrongly sortet rows in DOK";
select *
from test.SortDok_wrong;


Sincerely
Rune Runnestoe