Re: SQL question

Alexander Leites wrote: > > Hello, Matt. > > >select fname,count(*) from customer > >group by fname > >where count(*) > 1> [SNIP] > Try following: > > select fname,count(*) from customer cust_tbl1 > where (select count(*) from customer cust_tbl2 > where cust_tbl1.fname = cust_tbl2.fname) > 1 > group by fname> Better: select fname, count(*) from customer group by fname having count(*) > 1; This is MUCH faster than a correlated subquery, especially that one). (BTW the syntax guide shows HAVING clauses must precede GROUP BY clauses, the manual is wrong, GROUP BY comes first!) N.B.- ANY correlated subquery can be rewritten as a simple query or simple join! (Indeed ANY non-trivial SQL SELECT can be written two other ways! -- Kagel's Law) Art S. Kagel