Comments on: Is plan_hash_value a final say?

By: Subquery Factoring « Oracle Scratchpad

Subquery Factoring « Oracle Scratchpad — Thu, 16 Feb 2012 17:04:13 +0000

[...] is going to be different from the plan_hash_value for the other two (the plan hash value is derived from the line numbers, operations, and object [...]

By: PLAN_HASH_VALUE Limitations – PQ_DISTRIBUTE & skew in data distribution across parallel slaves « Oracle

Mon, 09 Aug 2010 21:04:18 +0000

[...] Filed under: Uncategorized — ksrivenu @ 12:04 pm Recently i read an article titled “Is plan_hash_value a final say?” on a Riyaj’s blog. As usual in his lucid style, Riyaj has demonstrated the shortcomings of the [...]

By: orainternals

orainternals — Fri, 06 Nov 2009 05:09:37 +0000

Hello Stellios
Thanks for visiting my blog. I corrected your comment and removed t
You are correct about “explain” command, not bind peeking (at least as of version 10g). But, point of blog is that filter predicates and access predicates can change without a change to plan_hash_value. This lead to a incorrect diagnosis and that’s exactly what I blogged about.

Cheers
Riyaj

By: Stellios

Stellios — Fri, 06 Nov 2009 04:38:34 +0000

I’ve tried posting this three times now – site keept reporting “discarded”.

I think your test case may be incorrect. Using autotrace and EXPLAIN for SQL that contains bind variables may not necessarily produce the correct execution plan. EXPLAIN does NOT perform bind variable peeking and is especially prone to being incorrect when indxes are involved.

Use

SELECT * FROM TABLE (DBMS_XPLAN.DISPLAY_CURSOR(‘&SQL_ID’)) (although you can use PLAN_HASH_VALUE also – documentation does not say this).

Also see Why EXPLAIN PLAN should not be used with Bind Variables

By: Blogroll Report 11/09/2009 – 18/09/2009 « Coskan’s Approach to Oracle

Blogroll Report 11/09/2009 – 18/09/2009 « Coskan’s Approach to Oracle — Mon, 21 Sep 2009 15:58:55 +0000

[...] Riyaj Shamsudeen – Is plan_hash_value a final say? [...]

By: orainternals

orainternals — Sat, 19 Sep 2009 04:14:56 +0000

I just read Randolf post. Very interesting indeed.
Thanks for posting here.

By: coskan

coskan — Fri, 18 Sep 2009 17:55:41 +0000

I think for further reading these two from Randolf Geist should be added

http://oracle-randolf.blogspot.com/2009/07/planhashvalue-how-equal-and-stable-are.html

http://oracle-randolf.blogspot.com/2009/07/planhashvalue-how-equal-and-stable-are_26.html