Comments for Pete Scott's random notes

Comment on Where to buy your data warehouse in 2010 by SoftwareTribes Teradata

SoftwareTribes Teradata — Tue, 27 Oct 2009 14:06:20 +0000

Nice comparison between Teradata and netezza

http://blog.pxltd.ca/2006/06/netezza_vs_tera.html

Comment on Hunting deadlocks part 2 by Ask The Real Tom » Oracle Deadlock ORA-00060 - Wait for Shared Lock

Ask The Real Tom » Oracle Deadlock ORA-00060 - Wait for Shared Lock — Mon, 12 Oct 2009 13:11:51 +0000

[...] support) did not help, oracle support could not help. The nessesary infomation if found here: http://pjsrandom.wordpress.com/2006/02/28/hunting-deadlocks-part-2/ http://oracledoug.com/serendipity/index.php?/archives/1014-Being-Open-minded.html Solution: The [...]

Comment on Hunting deadlocks part 2 by oracle tom

oracle tom — Mon, 12 Oct 2009 11:06:21 +0000

Gead Blog.
You saved me after 3 weeks searching. Oracle Support insited on a classic row lock an could not tell me why it’s waiting for shared lock.

Comment on UNGROUP BY by Paul Mclachlan

Paul Mclachlan — Thu, 09 Apr 2009 06:30:38 +0000

Joking aside, we basically have this feature, although using it for sales would be a bit… silly.

A more practical example is what we use it for — activity costing/accounting. So, if you’re spending $400k on a data-center and you want to know what the data-center is being used for, one way you could work this out — from a decision making perspective, is pull another feed, such as a list of servers per datacenter, and weight the data-center cost out per-server based on some criteria — for example, a combination of the physical size of the server, how much power it uses, etc.

It isn’t very much more difficult than a single statement along the lines of the above — a kind of “WEIGHT BY Server.Size+Server.Power/1500″ type of thing. Then all the usual BI reporting and analysis is available on top of the ending model.

Comment on UNGROUP BY by David Aldridge

David Aldridge — Thu, 02 Apr 2009 03:51:31 +0000

Comment on UNGROUP BY by Mark

Mark — Wed, 01 Apr 2009 10:49:44 +0000

Now I would have run this on the Rittman Mead blog, and actually said it was a secret new feature in 11gR2…

Comment on Technical blogs by mark modricker

mark modricker — Mon, 19 Jan 2009 23:41:19 +0000

Pete,

Hey there. I am Vp of Sales and Marketing for ALTA Information Services. We are launching our new Corporate website this week, wondered if you could be so kind as to post our link on your website?

the URL is http://www.altainfoserv.com

Let me know!

Thanks!

Mark Modricker

Comment on Athens (Greece) by William Robertson

William Robertson — Tue, 25 Nov 2008 07:53:53 +0000

Yes we should have got more done before 1848. If only we’d had a Renaissance or something.

Comment on Techniques to reduce IO – partitioning and compression by Peter Scott

Peter Scott — Tue, 29 Jul 2008 16:29:42 +0000

Sometimes you want to arrange data into partitions based on a key that does not fit into ranges for example countries could be partitioned by continent, but there are no simple algorithm to say that Belgium is Europe and Borneo is in Asia. In cases like this we can use list partitions to define which known value of partition key fits into which partition. This can give great performance for queries that select many key values from the same partition – for my example we could look at sales in Europe and only look at the Europe partition. But one downside of list partitions is that the key must be something that does not move about a lot – countries rarely change continent but people can change town, so basing a list on a volatile attribute is not such a good idea

Comment on Techniques to reduce IO – partitioning and compression by liyanq

liyanq — Mon, 28 Jul 2008 05:34:27 +0000

I like you website so much…
Learning from you…
could you please give some sample on how list patitioning on DW project?

thanks so much,
liyan