Archives of the TeradataForum

Message Posted: Wed, 10 Feb 2016 @ 09:36:02 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

You can try:

     sel grp, count(*) cnt,  min(keyval),max(keyval) from (
       sel keyval , quantile(3,keyval)  grp from tst
      ) dq
     group by 1 order by 1

or - if you want to keep the ansi standard:

     sel grp, count(*) cnt,  min(keyval),max(keyval) from (
       sel keyval , (RANK() OVER (ORDER BY keyval) - 1) * 3 / COUNT(*) OVER() grp from tst
      ) dq
     group by 1 order by 1

	<Prev	Next>		<<First	<Prev	Next>	Last>>