Re: Daily SQL Quiz

In article <1993May5.085551.10388@sei.cmu.edu> marc@sei.cmu.edu (Marc Graham) writes: > >It may be that even writing SQL code is more fun than working. (I >guess that only applies if your not being paid to write SQL code.) >Anyway, the problem is to > > select the row from a table having the largest date > field among all rows of that table having the same task_Id. > >The solution proposed by Aahz (whose .signature is to be commended) is >in the right direction but not at the proper destination. What is >needed here is a subquery to find the value of the date. Comme ca: > >select whatever_you_want >from Table Outer >where date_field = > (select max(date_field) > from Table Inner > where Inner.task_id = outer.task_id);> >Inner and Outer are called corelation names (but don't ask me why >they're called that). That's "correlation" name and it comes from the concept that it correlates the rows from the subquery with the rows in the outer query. This is why a subquery that has outer references is called a correlated subquery (whereas one that has no outer references is merely a subquery). This was the initial purpose for the use of a correlation name in a from clause. It is also used to distinguish a table from itself when a self-join is done. Jon Rosen