Re: more anti-postgresql FUD

On 10/13/06, Roman Neuhauser <neuhauser(at)sigpipe(dot)cz> wrote:
> # mmoncure(at)gmail(dot)com / 2006-10-10 14:16:19 -0400:
> > FUD from another open source project is really poor form, particulary
> > when not in competing segements where a little bit of competitive
> > rivalry is expected.
>
> OMG WTF what FUD???

please see my later comments. 'fud' is not a great term. however, if
you are going to publish remarks about another project that might be
perceived as disparaging, please keep them up to date and factually
relevant. I can write queries that are 10x slower on mysql that
postgresql but that ultimately means nothing. the major point thought
is that zabbix does *not* run 10x slower on postgresql and I am going
to prove it.

btw, i never said anything disparaging about mysql or zabbix. i am
focused like a laser beam on the comments in the documentation and the
greater implications for the community.

> I had a large (several milion rows), indexed table, same data, in
> MySQL (4.0.x) and PostgreSQL (late 7.4), on the same RHEL or FreeBSD
> (don't remember) machine. Walking over the table with
>
> SELECT * FROM TABLE ORDER BY pk LIMIT 10 OFFSET N;

using offset to walk a table is extremely poor form because of:
* poor performance
* single user mentality
* flat file mentality

databases are lousy at this becuase they inheritly do not support
abolute addressing of data -- nore should they, beause this is not
what sql is all about. in short, 'offset' is a hack, albeit a useful
one in some cases, but dont gripe when it doesn't deliver the goods.

for server side browsing use cursors or a hybrid pl/pgqsl loop. for
client side, browse fetching relative to the last key:

select * from foo where p > p1 order by p limit k;

in 8.2, we get proper comparisons so you can do this with multiple part keys:

select * from foo where (a1,b1,b1) > (a,b,c) order by a,b,c limit k;

for fast dynamic browsing you can vary k for progressive fetches.

> or the MySQL equivalent, MySQL was several times faster than
> PostgreSQL, but the times were getting longer and longer....
> As N grew in increments of 10, it took ages for MySQL to return
> the rows. PostgreSQL... Well, it was as "slow" with N=100000 as it was
> with N=0.
>
> > * MySQL is used as a primary development platform.
>
> How does *this* qualify as FUD? Or are *you* spreading FUD to scare
> people from even mentioning the software?

I think zabbix is fine software. I would hopefully prefer that if
someone were to write what could be perceived as negative things about
postgresql, they would back it up with facts better than 'update foo
set id = 0' ran a million times or 'select * from foo limit 1 offset
100000'

> I don't like MySQL. I hate it when people put cheerleading where reason
> should prevail.

outside of the 'fud' statement, which was a hastily written reaction,
my tone has been more constructive criticism.

merlin