Query result is very slow

Question

i have two table master detail, for instance dept and emp
the situtation is this. i need the recod of those departements who has clerks in specific time hired. i wrote following query. but due to millions of records in two tables its very very slow.
select * from emp a,dept b
where a.deptno in (select distinct deptno from emp where job='clerk')
and hiredate >= '01-jun-2004' and hiredate <='01-jan-2007'
and a.deptno=b.deptno
can any body tune it.

Answer 1

One thing I am seeing, that I find very troubling, is that posters such as the OP on this thread seem to be facing these tasks in a classroom or testing environment where they don't actually have access to Oracle and SQL*Plus.
They are being asked to tune something that exists only on paper and don't have the tools to provide an explain plan, probably don't even have the ability to do a describe on the table(s) because they don't actually exist.
If this is the case then the education system is cheating these people because tuning is not done in a room with the lights off by guessing. It is done, and only done, with:
EXPLAIN PLAN FOR
DBMS_XPLAN
TKPROF
a look at the statistics, etc.
Here's a real-world example of what is wrong with what I see. Anyone want to guess which of these queries is the most efficient?
SELECT srvr_id
FROM servers
INTERSECT
SELECT srvr_id
FROM serv_inst;
SELECT srvr_id
FROM servers
WHERE srvr_id IN (
  SELECT srvr_id
  FROM serv_inst);
SELECT srvr_id
FROM servers
WHERE srvr_id IN (
  SELECT i.srvr_id
  FROM serv_inst i, servers s
  WHERE i.srvr_id = s.srvr_id);
SELECT DISTINCT s.srvr_id
FROM servers s, serv_inst i
WHERE s.srvr_id = i.srvr_id;
SELECT /*+ NO_USE_NL(s,i) */ DISTINCT s.srvr_id
FROM servers s, serv_inst i
WHERE s.srvr_id = i.srvr_id;
SELECT DISTINCT srvr_id
FROM servers
WHERE srvr_id NOT IN (
  SELECT srvr_id
  FROM servers
  MINUS
  SELECT srvr_id
  FROM serv_inst);
SELECT srvr_id
FROM servers s
WHERE EXISTS (
  SELECT srvr_id
  FROM serv_inst i
  WHERE s.srvr_id = i.srvr_id);
WITH q AS (
  SELECT DISTINCT s.srvr_id
  FROM servers s, serv_inst i
  WHERE s.srvr_id = i.srvr_id)
SELECT * FROM q;
SELECT DISTINCT s.srvr_id
FROM servers s, serv_inst i
WHERE s.srvr_id(+) = i.srvr_id;
SELECT srvr_id
FROM (
  SELECT srvr_id, SUM(cnt) SUMCNT
  FROM (
    SELECT DISTINCT srvr_id, 1 AS CNT
    FROM servers
    UNION ALL
    SELECT DISTINCT srvr_id, 1
    FROM serv_inst)
  GROUP BY srvr_id)
WHERE sumcnt = 2;
SELECT DISTINCT s.srvr_id
FROM servers s, serv_inst i
WHERE s.srvr_id+0 = i.srvr_id+0;And yes they all return the exact same result set using the test data.
The chance that anyone can guess the most efficient query looking at what I just presented is precisely zero.
Because the query most efficient in 8.1.7.4 is not the most efficient query in 9.2.0.4 which is not the most efficient query in 10.2.0.2.