Re: A test to add to the crashme test

From: Michael Robinson <robinson(at)netrinsics(dot)com>
To: robinson(at)netrinsics(dot)com, tgl(at)sss(dot)pgh(dot)pa(dot)us
Cc: pgsql-hackers(at)hub(dot)org
Subject: Re: A test to add to the crashme test
Date: 2000-05-22 07:46:39
Message-ID: 200005220746.PAA33389@netrinsics.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> writes:
>> P.S. When it comes to ROLAP, though, MySQL kicks PostgreSQL's butt. For that
>> application only, I use MySQL.
>
>Er ... "ROLAP"? Expound, please.

Relational On-Line Analytical Processing. As opposed to Multidimensional
Online Analytical Processing (MOLAP), the other kind of OLAP.

The basic principle of operation is that you put all your data in a big
star (or snowflake) schema, and then pare down your "cube" by pre-aggregating
various dimensions of interest into various auxillary tables.

It works much better than MOLAP for big, sparse, high-dimensional data
(like, for example, six months of log data from an active e-commerce/content
website).

MySQL is extremely well suited for it: the data is essentially "read-only"
so transactions, locking, etc., are not an issue, the per-row overhead is
extremely small (important when you have hundreds of millions of short
records), and the speed, especially with prudent indexing and datatype
selection, is scorching fast.

Just don't ever put any data in it that you can't reconstruct from scratch.

-Michael

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Hannu Krosing 2000-05-22 08:16:07 Re: Last call for comments: fmgr rewrite [LONG]
Previous Message Kaare Rasmussen 2000-05-22 07:37:04 rolap