A parallel GROUP BY clause can generate incorrect results for variables with non-standard length columns if a HAVING clause is used.

The following code illustrates a case that can produce incorrect results in columns YEAR and SUM(PAY_MTS):

 
  libname foo sasspds 'public' server=hostname.5400 user='anonymous' 
  ip=yes libgen=yes;

  data foo.mytest;
    length year 3 abc 3 paid $1 chargeable $1 pay_mts 8;
    input year abc paid $ chargeable $ pay_mts;
    cards;
    2010 100 y n 200
    ;
  run;

  proc sql;
  select distinct abc, chargeable, paid, year, sum(pay_mts)
     from foo.mytest
     group by abc, chargeable, paid, year having count(*) gt 0;
  quit;

Select the Hot Fix tab in this note to access the hot fix for this issue.

Operating System and Release Information

A fix for this issue for Scalable Performance Data Server 4.52 is available at:

A fix for this issue for SPD Server 4.4 is available at: