RE: USE_NL with or without ORDERED

From: Shevtsov, Eduard <EShevtsov_at_flagship.ru>
Date: Mon, 17 Sep 2001 02:38:35 -0700
Message-ID: <F001.0039016E.20010917021020@fatcity.com>

Jonathan,

If you haven't droped your test tables yet could you try /*+ USE_NL(c e) */ Yeah, sounds strange, but...

SQL> select /*+ use_nl(c e) */

  2          e.first_name
  3   ,      e.last_name
  4   ,      c.short_name

  5 from courses c,employees e
  6 where e.emp_id = c.dev_id
  7 /

1018 rows selected.

Execution Plan

---

   0 SELECT STATEMENT Optimizer=CHOOSE (Cost=1022 Card=1018 Bytes

          =31558)  

   1 0 NESTED LOOPS (Cost=1022 Card=1018 Bytes=31558)

   2 1 TABLE ACCESS (FULL) OF 'COURSES' (Cost=4 Card=1018 Bytes

          =14252)  

   3 1 TABLE ACCESS (BY INDEX ROWID) OF 'EMPLOYEES' (Cost=1 Car

          d=15132 Bytes=257244)  

   4 3 INDEX (UNIQUE SCAN) OF 'EMP_PK' (UNIQUE) SQL> select /*+ use_nl(c e) */

  2          e.first_name
  3   ,      e.last_name
  4   ,      c.short_name

  5 from employees e,courses c
  6 where e.emp_id = c.dev_id
  7 /

1018 rows selected.

Execution Plan

---

   0 SELECT STATEMENT Optimizer=CHOOSE (Cost=1022 Card=1018 Bytes

          =31558)  

   1 0 NESTED LOOPS (Cost=1022 Card=1018 Bytes=31558)

   2 1 TABLE ACCESS (FULL) OF 'COURSES' (Cost=4 Card=1018 Bytes

          =14252)  

   3 1 TABLE ACCESS (BY INDEX ROWID) OF 'EMPLOYEES' (Cost=1 Car

          d=15132 Bytes=257244)  

   4 3 INDEX (UNIQUE SCAN) OF 'EMP_PK' (UNIQUE) Regards,
Ed

>
>
>
> Larry's answer is totally correct -
>
> The SQL reads:
> select /*+ use_nl(t2) *./
> etc
> from t1,t2
> where t1.id = t2.id
> etc
>
> But manages to hash from T1 to T2.
>
> (I had to play around a little bit to
> make the tables and rows the right
> size for this to happen).
>
> Looking at the 10053 trace, I got the lines:
> (with lots of cuts and editing)
>
> General Join Order:
> Join Order 1 T1 -> T2
> use_nl
> calculations - cost = 92
> NO LINES for sort/merge
> NO LINES for hash join
> so the hint IS constraining Oracle to
> consider only NL when joining T2.
>
> Join Order 2 T2 -> T1
> Use_nl
> calculations - cost = 127
> sort/merge
> calculations - cost = 64
> hash join
> calculations - cost = 47
> (sides swapped)
>
> So Oracle arrives at the best path by
> considering the join order T2 -> T1,
> which makes the hint ignorable, but
> then swaps the table order at execution
> time, to produce the apparent contradiction.
>

-- 
Please see the official ORACLE-L FAQ: http://www.orafaq.com
-- 
Author: Shevtsov, Eduard
  INET: EShevtsov_at_flagship.ru

Fat City Network Services    -- (858) 538-5051  FAX: (858) 538-5051
San Diego, California        -- Public Internet access / Mailing Lists
--------------------------------------------------------------------
To REMOVE yourself from this mailing list, send an E-Mail message
to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from).  You may
also send the HELP command for other information (like subscribing).

Received on Mon Sep 17 2001 - 04:38:35 CDT