This probably isn't a very popular question here, but I'll ask it
nontheless...

In the past years, I've run into remarks about MySQL not being a DBMS
to take seriously by some developers. I've been a MySQL user for quite
some time now and I have trouble figuring out why some developers are
so much against MySQL. While I'm beginning to believe part of these
remarks are true, but only to a small degree, I'm wondering about the
real reasons as to why some still think MySQL is inferior.

I concurr that I found check constraints to be missing and BOOLEANs to
constrain to an INTEGER in the range of a TINYINT instead of to TRUE
and FALSE. Additionally there are some CURRENT_TIMESTAMP issues that I
don't like. But are these enough arguments to tout MySQL as an
inferior or "not a real" DBMS? What are the real reasons here - if
they exist?

Karsten

On 12/20/2010 5:19 PM, Karsten Wutzke wrote:
There are a lot of problems with MySQL - performance being one of them.
MySQL is OK - but has nowhere near the performance required for many
databases. Also, not supporting SQL standards (i.e. recursive SQL) is a
big one.

I have never seen any benchmark where MySQL will match or outperform any
of the other databases, and in most it is significantly worse.
Additionally, the other databases pretty much fully support the latest
SQL standards. This is especially true if you get into the InnoDB
engine, which is much slower than ISAM.

Not to say MySQL isn't worth using - it is fine for low activity web
sites, for instance. But you won't find many companies using it for
critical data on high-traffic servers.

--
==================
Remove the "x" from my email address
Jerry Stuckle
JDS Computer Training Corp.
jstucklex (AT) attglobal (DOT) net
==================

"Jerry Stuckle" <jstucklex (AT) attglobal (DOT) net> wrote

Isn't that a little misleading as the main vendors (Oracle, Microsoft etc)
often have it written into their terms and conditions that benchmarking
figures may not be published. Hence why its difficult to find any
benchmarking figures comparing two or more databases.

I believe wikipedia still uses MySQL.
--
Brian Cryer
http://www.cryer.co.uk/brian

"Karsten Wutzke" <kwutzke (AT) web (DOT) de> wrote

There was a time when MySQL didn't support transactions and that alone would
have been a good reason to give it a miss for any serious system.
Unfortunatly it can take a while for that type of "stigma" (for what of a
better word) to be overcome (even when its no longer valid).

Whilst I fully acknowledge that MySQL is still lacking in some areas, the
main reason (IMHO) that some people still look down on it are simple human
factors: ignorance, bias (under which I'm including existing skill sets) and
snob-value. I'm not saying that MySQL should always be the database of
choice, its always worth considering fitness for purpose and your own
benchmarking if you have time. Google for example doesn't use MySQL (nor I
think does it use Oracle or SQL Server), but it seems perfectly adequate for
wikipedia.

Personally I wouldn't worry about some developers not taking MySQL seriously
as a DBMS, I've worked with some developers who couldn't be taken seriously.
Better to worry about whether or not MySQL is suitable (or best) for your
intended use.
--
Brian Cryer
http://www.cryer.co.uk/brian

Karsten Wutzke <kwutzke (AT) web (DOT) de> wrote:
Times are changing and so are requirements for DBMSes. Many developers
seem to stick to what they have learned how a DBMS should work. That's
basically Codds ACID rules. Such people cannot understand that for some
use cases you just don't need triggers or stored routines or even
referential constraints. A typical example of such a guy is Jerry S.

Of course there *are* applications that can profit from all those
features. But there are fewer of these every day. It's the web and the
cloud that changes everything.

If you look at current trends in databases, then many businesses move
*away* from monolithic RDBMS architectures. NoSQL, CouchDB, Hadoop,
$YOU_NAME_IT - those are as far from MySQL as MySQL is from i.e. DB2.
The approach is lightweightedness, horizontal scalability, distribution.

MySQL itself is on the edge. I would even say this edge has cut
already. Some years ago (about with the 5.0 release) MySQL did a huge
step towards the oldfashioned, traditional DBMSes. It added a lot of
advanced features. And only a little later there was the splitting-off
of the Drizzle project. While MySQL seems to drift towards the heavy
traditional DBMSes, Drizzle moves in the other direction.

Still MySQL is flexible enough to work for both groups of people. For
the traditionalists there are InnoDB, sql_mode="traditional" (sic!)
and all the bells and whistles added in 5.x. For the other guys there
is still MyISAM, soon also Maria, replication, etc.

Finally: anybody who understands German (or trusts Google translate
should read Kris:

http://blog.koehntopp.de/archives/29...-Postgres.html
(this was also posted as comment to the Heise Ticker)


XL

"Brian Cryer" <not.here (AT) localhost (DOT) invalid> wrote:

Exactly. "Know your tools"
Even the idea of a DBMS that is best choice for any application
seems strange to me.

Wrong. Google does use MySQL a lot, but of course not everywhere.
Look at the MySQL bug tracking system for bugs reported (and fixes
delivered) by Google. Or check the Google patches for MySQL.


XL

Jerry Stuckle <jstucklex (AT) attglobal (DOT) net> wrote:
This proves only one thing: your ignorance.

http://www.mysqlperformanceblog.com/...abase-contest/

Yes, this was 4 years ago. And of course this benchmark was unfair
because it played in a field where MySQL is strong and traditional
DBMSes are weak: web applications.

I really don't see the connection between "Standard conformance" and
"InnoDB slower than ISAM". I only quote you for your second claim.

In fact InnoDB has become much better and in general performs about
as good as MyISAM. There are edge cases where InnoDB is 10x faster.
Rule of thumb: the more writes (relatively to reads) you have, the
better InnoDB will be.

Yeah. Just Wikipedia. Or Facebook. Low activity sites, my ass!
YOU. KNOW. NOTHING!


XL

"Axel Schwenke" <axel.schwenke (AT) gmx (DOT) de> wrote

My comment was directed at how they manage their search indexes (for what of
a better term). I hadn't given any thought to what they might use for other
things - where I stand corrected. Thank you.
--
Brian Cryer
http://www.cryer.co.uk/brian

On 12/21/2010 4:43 AM, Brian Cryer wrote:
That doesn't stop the benchmarks from occurring.

So? Wikipedia traffic is high for web sites, but not for business sites.
--
==================
Remove the "x" from my email address
Jerry Stuckle
JDS Computer Training Corp.
jstucklex (AT) attglobal (DOT) net
==================

On 12/21/2010 6:30 AM, Axel Schwenke wrote:
Wow. MySQL/PHP vs., DB2/Java, Oracle/Java and PostreSQL/PHP. Apples
and Oranges. Also combinations absolutely used for all business
applications. And 3664 orders PER MINUTE! Gee, Impulse Power, Mr. Scott!

Now lets stack that up against REAL databases - like DB2/COBOL, Oracle/C
and similar - which run 100's of K transactions PER SECOND.

As I said - MySQL is OK for web applications. But it's not a serious
business database. But you wouldn't know because you've never seen a
real database or worked in the real world.

If you want things like foreign key constraints, you have to use InnoDB,
not MyISAM.

Which only shows how inefficient MyISAM is in writing data. But it
matters not - in the real world, the vast majority of database
operations are reads, not writes.

Low activity databases, yes. But you've never seen a heavily used
databases. I've seen large banks which run 100's of K transactions (not
reads/writes) per second. And airlines run in the millions of
transactions/sec.

MySQL isn't even in the same league as the big boys.


--
==================
Remove the "x" from my email address
Jerry Stuckle
JDS Computer Training Corp.
jstucklex (AT) attglobal (DOT) net
==================