Comments on: MySQL Query Performance Tips

By: code reduction

code reduction — Fri, 12 Mar 2010 09:47:47 +0000

Thanks very much for taking your time to create this very useful infos

By: Daily Links | AndySowards.com :: Professional Web Design, Development, Programming, Hacks, Downloads, Math and being a Web 2.0 Hipster?

Sat, 28 Feb 2009 01:07:54 +0000

[...] MySQL Query Performance Tips | 84 Bytes More awesome tips for MySQL (tags: tutorial optimization performance mysql optimisation) [...]

By: Artem Russakovskii

Artem Russakovskii — Fri, 27 Feb 2009 20:02:04 +0000

The last one “INSERT … ON DUPLICATE KEY update (INSERT IGNORE) to avoid having to SELECT” in some cases can be really bad for your slaves, as INSERT IGNORE will propagate to all slaves, causing slave lag. If you expect lots of INSERTs to dupe, it’s better to SELECT and INSERT selectively.

Artem

By: TravisO

TravisO — Mon, 03 Nov 2008 21:37:03 +0000

I’d also like to add, in extreme niche scenarios, using a Unix epoch style timestamp (with an INT datatype) is more efficient than the DateTime field. I can’t vouch for the performance difference, it’s probably pretty small, but back in 2001 with MySQL 4, I saw performance gains in the 8 fold difference (that’s 800%).

But I have faith in today’s hardware and MySQL 5 builds and I haven’t attempted to recreate my old benchmarks. Despite, the fact is MySQL has to store date/time internally, the format you see is massaged for humor consumptions, so conversions are being executed here, the less conversions the better. And on that note, I recently ran across a system where the coder irresponsibly retrieved all data for a certain date by doing

SELECT * FROM table WHERE date LIKE ’2008-11-01%’

Instead of simply using BETWEEN, obviously this incurred massive slowdowns.

By: peng wang

peng wang — Wed, 15 Oct 2008 12:34:37 +0000

excellent tips, very useful for newbie like me, thks for sharing ^_^