Roman "Dagi" Pichlik created a CZJUG channel on YouTube, so from now on all the videos won't be uploaded to Mediacast (which will be decomissioned on July 29) but to this new channel. I'm not able to compare these two services, but I perceive this as a step in the right direction. However there are several small issues I'm trying to solve ...
We all know than once a project passes the initial hectic phase when the design is tuned (and before it reaches the the phase "we have to deliver it to the customer and it's not ready"), the development is a rather boring experience. Occasionally you'll learn something new, you'll solve some unexpected problem, but in general it's quite monotonous work. Some vendors obviously are aware of this "issue" and obviously decided to take heed to make the development much more fun, as their representatives obviously think that bore is a killer disease.
On this friday night (May 21) a local version of "PostgreSQL testing night," an overnight meeting of PostgreSQL users with the aim of testing functionality of the new version (currently 9.0), took place in Prague. Six enthusiasts participated - Pavel Stehule, Honza Matousek, Vaclav Novotny, Tzvetan Tzankov, Petr Michalek and me.
We all know there are fundamental differences between relational and object model - called "object-relation impedance mismatch." If we are going to build relational object model on top of relational database (which is natural if using object-oriented language), we have to solve that difference somehow, and ORM tools are one of the options. But are the often mentioned benefits real?
This puzzler is not a clean SQL puzzler, as it is related not only to the SQL function now(), but to a PHP function time() too. Are you accessing PostgreSQL from PHP? In that case you should solve this puzzler easily ;-)
As I became responsible for recording and processing videos from regular CZJUG meetings, I had to learn how to perform basic video processing in Linux - years ago I did this on Windows in Adobe Premiere, so I have been looking fro a similarly comfortable and flexible solution. I guess you are trying to solve about the same problem, so lets review my solution based on ffmpeg and kdenlive.
Installing Oracle 11g (R2) on an unsupported distribution (currently there are three supported ones - Oracle Enterprise Linux, Red Hat and SuSe) may be a bit of a problem - especially if the distribution is "progressive" i.e. contains fresh versions of different tools (compiler, libraries) and is not based on RPM packages. And both of these are true for Gentoo, so lets see what problems may occur during the installation (and how to solve them).
If you plan to partition a continually growing table to parts of a limited size, you'll have to find a suitable way to create new partitions. It's not advisable to do that manually (at least on a production system), it's much better to do that automatically. Lets see several possible solutions ...
In PostgreSQL there are two different ways to perform partitioning - using rules and using triggers. One of the usual reasons why a partitioningu is used, is a performance, so lets analyze differences between these two ways regarding performance.
In contrast to the previous articles focused on technology, this one is a short reasoning about hints in Oracle. My approach to hints is somehow ambiguous - from time to time I do admit hints may be useful, but most of the time I just hate them ...