Comments on: SQL Speed Test: IN vs OR http://www.webmastersbydesign.com/2008/07/13/sql-speed-test-in-vs-or/ Christian Webmaster Resources Sun, 28 Feb 2010 13:34:31 +0000 http://wordpress.org/?v=2.9.2 hourly 1 By: Dmitri P http://www.webmastersbydesign.com/2008/07/13/sql-speed-test-in-vs-or/comment-page-1/#comment-225 Dmitri P Tue, 15 Jul 2008 06:45:16 +0000 http://www.webmastersbydesign.com/?p=84#comment-225 I agree that smaller samplings can be anecdotal.. Did the clock tick over a minute, did you move the mouse, etc. 10,000 - 100,000 rows would give you a repeatable result. It seems like this test weighs extra text on the wire (OR) against processing power.. DB extrapolation (IN). It might be fun to pursue this further down the path. Good article regardless. -dp I agree that smaller samplings can be anecdotal.. Did the clock tick over a minute, did you move the mouse, etc. 10,000 – 100,000 rows would give you a repeatable result.

It seems like this test weighs extra text on the wire (OR) against processing power.. DB extrapolation (IN).

It might be fun to pursue this further down the path. Good article regardless.

-dp

]]> By: anonymous http://www.webmastersbydesign.com/2008/07/13/sql-speed-test-in-vs-or/comment-page-1/#comment-220 anonymous Mon, 14 Jul 2008 14:28:47 +0000 http://www.webmastersbydesign.com/?p=84#comment-220 Did you run these tests immediately following each other? If you had the reduced time could be from the RDBMS having already loaded its cache. Did you run these tests immediately following each other? If you had the reduced time could be from the RDBMS having already loaded its cache.

]]> By: RMagnus http://www.webmastersbydesign.com/2008/07/13/sql-speed-test-in-vs-or/comment-page-1/#comment-215 RMagnus Mon, 14 Jul 2008 01:27:08 +0000 http://www.webmastersbydesign.com/?p=84#comment-215 I am curious to to see the statistics (ie: SQL Explain) that the DB Engine would report about the different queries. Having worked around database engines for several years*, I have seen a dramatic change in how the DB engines interpret, optimize, and execute the SQL we hand it. Only a few years ago, it seemed it was the burden of the programmer to implement efficient SQL code, but now the engines we use now are amazingly smarter, and are able to "re-interpret" the SQL we hand it. So I am wondering how Oracle engine above really interpreted your SQL. I am actually surprised from your findings, since in my experience "OR" branching always seems to be the longest path to follow. I am curious to to see the statistics (ie: SQL Explain) that the DB Engine would report about the different queries.

Having worked around database engines for several years*, I have seen a dramatic change in how the DB engines interpret, optimize, and execute the SQL we hand it.

Only a few years ago, it seemed it was the burden of the programmer to implement efficient SQL code, but now the engines we use now are amazingly smarter, and are able to “re-interpret” the SQL we hand it. So I am wondering how Oracle engine above really interpreted your SQL.

I am actually surprised from your findings, since in my experience “OR” branching always seems to be the longest path to follow.

]]>